Quality, value, satisfaction, trust, a

Volume 2 

Contents: 

A Global Examination Of Relationship Marketing Concepts In A Retail Banking Context ....3 

Philippe AURIER, University Montpellier 2 

Gilles N'GOALA, EDHEC Business School 

A study of the impact of shopping orientation and gender on the value-satisfaction link during 

an electronic catalog visit.................................................................................................................55 

Christine Gonzalez, Toulouse Business School 

The Influence of Intangibility on Perceived Evaluation 

Difficulty and Risk: Brand and 

Generic Product-Category Perspectives 

........................................................................................79 

Michel 

Laroche, Concordia University 

Maria 

Kalamas, Concordia University 

Gordon H. G. McDougall, Wilfrid Laurier University 

Filip Bartos, Concordia University 

Yi Zhong, Concordia University 

Exploring the WOW in Online Auction Feedback .....................................................................117 

Bruce D. Weinberg, Bentley University 

Lenita 

Davis , The University of Alabama 

Physical Behaviour 

in Stores and Spatial Configuration: Two Exploratory Studies..............145 

Gaël 

Bonnin, EDHEC Business School 

Asymmetric Quality Tier Competition: An Alternative Explanation. ......................................170 

K. Sivakumar 

Modeling The Impact Of Internet Atmospherics On Surfer Behavior.....................................191 

Marie-Odile Richard, École des Hautes Études Commerciales 

Mapping Retailing Enterpreneur Decisions 

And Behavior .......................................................231 

Arch G. Woodside, Boston College 

1

An Empirical Examination of Retail Convenience for In-Store and Online Shoppers...........251 

Nicole Ponder, Mississippi State University 

Michelle 

Bednarz, Mississippi State University 

What influences customer reaction to incompatibility? An empirical investigation...............281 

2

A Global Examination Of Relationship Marketing Concepts In A Retail Banking Context 

Philippe AURIER, University Montpellier 2 

Gilles N'GOALA, EDHEC Business School 

Abstract. This paper tests the validity and the causal structure of the main concepts developed in the 

relationship marketing literature, including perceived quality (technical and functional), value, satisfaction, 

trust and commitment to the brand. Working with two samples each consisting of 2,150 customers of a large 

bank, we first study the unidimensionality of these concepts and underline their high correlations. We then 

study their causal structure and show that our data are compatible with a "relational chain" structure with 

significant direct effects of both quality components (technical and functional) on value, and effects of value 

on satisfaction, of satisfaction on trust, and finally, of trust on commitment. Indeed, all these links are 

positive, strong, and nearly equal in intensity (close to 0.5). 

Key Words. Relationship marketing, Quality, Value, Satisfaction, Trust, Commitment, Brand-Consumer 

Relationship, Service. 

3

INTRODUCTION 

In the course of the last twenty years or so, numerous studies focusing on commercial relations have 

demonstrated the importance of perceived quality, value, satisfaction, trust, and commitment in the formation 

of long-term customer relationships. Twenty years after the original article by Berry (1983), the conceptual 

framework of relationship marketing is now well established and can be easily applied to relationships 

between consumers and brands. Without claiming to provide an exhaustive summary, we might mention that 

the literature strongly emphasizes both consumer judgments of a product's perceived quality and value and 

consumer satisfaction in evaluating consumption experiences. Furthermore, the literature considers trust of, 

and commitment to, the brand as the main antecedents of loyalty (continuity/repurchase) and long-term 

cooperation. While no one disputes the theoretical utility of this research, there are nevertheless several 

questions that remain unresolved. 

First of all, different concepts that characterize the relationship to the brand have been developed separately: 

some of the studies focus on post-purchase processes, while others are centered on relationships of trust and 

long-term commitment. Quality, value, satisfaction, trust, and commitment have often been presented and 

measured individually, as separate features in the mind of the judging consumer; each concept is covered in a 

different body of literature. Yet, the many theoretical similarities between these concepts lead us to doubt 

their separate existence, or at least their discriminant validity. Taking into account the high correlations that 

must exist between these concepts, what then are their specific contributions to explaining consumer 

behavior? Furthermore, the causal structure of these concepts has not been approached in a overall fashion. 

While the links between quality and satisfaction and between trust and commitment have been studied 

extensively, to our knowledge no research has tested simultaneously the overall causal structure. Finally, at 

the empirical level, much of the research has been conducted on limited samples -- being either small or 

unscientifically constituted (using students or other convenient populations). As a result, the findings 

obtained from these studies lack both internal and external validity. 

This article tries to overcome these three limitations. Our goal is not to propose and validate a new theory of 

relationship marketing. We attempt instead to put the major concepts of relationship marketing in perspective 

4

and to measure them simultaneously using a random, heterogeneous sample of customers. After a brief 

review of the literature which will serve as a reminder of the concepts being examined (quality, value, 

satisfaction, trust, and commitment) and their assumed causal relations, we will study their trait validity, test 

a model of their causal structure, and then compare it to two alternative models emanating from the literature. 

We conducted our initial survey on a probabilistic sample of 2,150 customers of a large bank, matched with a 

similar sample (2,150 customers) taken from the same population and polled one year later. 

CONCEPTUAL FRAMEWORK 

With the advent of relationship marketing, the relationship between a consumer and an object (product, 

brand, logo, etc.) has become a popular topic of research. In the relationship marketing framework, the 

concept of "brand" is no longer limited to "extrinsic" features of a product, with which one associates 

different functional, emotional, or symbolic benefits. More broadly, a brand represents the anchor point of a 

lasting relationship that is built on past consumption experiences and those yet to come (Berry, 2000). 

Main Concepts Characterizing the Relationship to the Brand 

This brand – customer relationship is complex and cannot be summarized easily (Fournier, 1998). 

Nevertheless five concepts are frequently used in the relationship marketing literature: commitment, trust, 

satisfaction, perceived value and quality. They seem to be key concepts in cutomer-relationship 

understanding and they can hereafter be used to build a simplified and coherent model. Three of them 

(quality, value, and satisfaction) are rooted in evaluations of consumption experiences. These concepts can 

be considered from a transactional viewpoint (evaluation of a specific experience) or from a cumulative 

viewpoint (evaluation of the relationship). The other two concepts (trust and commitment), which represent 

more abstract judgments of the brand, are oriented toward the future and are reinforced or weakened by a 

succession of purchase and consumption experiences. 

Perceived quality: Perceived quality is generally considered the foundation of the relational process. It 

denotes "the consumer's judgment about a product’s overall excellence or superiority" (Zeithaml, 1988). 

Perceived quality is derived from the individual's evaluation of the intrinsic and extrinsic features of the 

object, based on initial expectations, in comparison with what competitors have to offer. In the service sector, 

perceived quality has been explored in several studies (Grönroos, 1983; Zeithaml, 1988; Parasuraman, 

5

Zeithaml and Berry, 1988; Zeithaml, Berry and Parasuraman, 1996; Brady and Cronin, 2001). Service, 

simultaneously produced, purchased and consumed, can be evaluated on the basis of the core service 

(technical quality) and of the interaction between the customer and the service personnel (functional quality) 

as Grönroos (1983) postulated. The concept of service also depends upon the interaction with the physical 

environment in which the service is provided (atmosphere, type of people who frequent the location), as 

Bitner (1990) showed. 

Perceived value: If perceived quality affects evaluation of benefits in comparison to offerings of competitors, 

these benefits must be considered in relation to the resources expended during the purchase-consumption 

cycle. The information search prior to purchase, the price paid, the time spent, and various efforts put forth 

during consumption, all constitute factors that reduce the value perceived by the consumer (Lai, 1995; Slater 

and Narver, 2000). Therefore, perceived value represents an appraisal of the exchange. In line with the 

approach developed in economics, consumer behavior is presented as a process of producing utility by 

applying monetary and temporal resources (Becker, 1976). Although other complementary sources of value 

exist (Holbrook, 1999), a product's quality or excellence remains the fundamental basis for value creation in 

the exchange (Zeithaml, 1988). 

Satisfaction: For Oliver (1997, p. 144), satisfaction denotes evaluation of the results of a consumption 

experience based on a set of goals or standards, which translates into fulfillment, under-fulfillment, or over- 

fulfillment. While the concepts of perceived quality and perceived value remain closely linked to the 

evaluated object (good or service), satisfaction describes the consumer's cognitive and affective states 

resulting from the consumption experience (Boulding, Kalra, Staelin, and Zeithaml, 1993; Oliver, 1993; 

Iacobucci, Grayson, and Ostrom, 1994). Indeed, perceived quality involves comparing performance to that of 

competitors, value involves a comparison against accepted sacrifices, while satisfaction is based on 

confirmation of individual expectations (predictive or normative standards). Satisfaction is not limited to 

cognitive evaluation, it involves an abstract state that is both cognitive and affective and that characterizes 

the individual's personal experience with a product or brand (Fournier and Mick, 1999). Finally, unlike 

quality, satisfaction requires a concrete experience with a product. Taking a relational approach, Johnson, 

Anderson, and Fornell (1995) introduced cumulative satisfaction, an "abstract construct that describes 

6

customers’ total consumption experience with a product or service". Cumulative satisfaction allows one to 

grasp, at a given time, the updated picture of an individual's cumulative experiences to date, both satisfactory 

and unsatisfactory. Although other explanatory factors can be found within the relationship marketing 

framework, perceived quality and perceived value are the two principal conveyors of consumer satisfaction 

(Johnson, Anderson, and Fornell, 1995). 

Trust: Consumer satisfaction acts as a risk reducer for future transactions. Each satisfactory experience 

generates trust and acts as a positive reinforcement. Thus, satisfaction enhances trust, i.e., the consumer's 

assumption that the brand, as a personified entity, promises to have a foreseeable performance that conforms 

to the consumer's expectations, and agrees to willingly maintain that standard over time. While satisfaction 

results principally from an assessment of an earlier concrete instance of consumption, trust is more of a bet 

on the future of the relationship, thus integrating the notion of risk. Trust is founded on the anticipated 

capacity of the brand to regularly satisfy consumer expectations (reliability, credibility, general reputation) 

and in a more affective sphere, it goes back to the brand's will -- assumed but not proven -- to avoid doing 

anything that might be detrimental to it (and thus to enhance perceived benevolence, equity, integrity; 

Ganesan, 1994; Ganesan and Hess, 1997). Trust in a brand is principally oriented toward the future 

relationship and is strengthened by a succession of satisfactory consumption experiences. As such, trust 

represents one of the main antecedents of long-term commitment to a brand. 

Commitment: Along with trust, commitment is a central concept in relationship marketing (Dwyer, Schurr, 

and Oh, 1987; Morgan and Hunt, 1994; Garbarino and Johnson, 1999). In marketing, Morgan and Hunt 

(1994) defined commitment to the relationship as "an exchange partner believing that an ongoing 

relationship with another is so important as to warrant maximum efforts at maintaining it; that is, the 

committed party believes the relationship is worth working on to ensure that it endures indefinitely". 

Research on commitment in marketing is based on older studies conducted in the area of human resources 

management. In particular, Mowday, Steers, and Porter (1979) defined commitment to the organization as 

the relative intensity of identification and involvement in a particular organization. Commitment here can be 

characterized by at least three factors: a strong belief and acceptance of the organization's goals and values, a 

willingness to expend considerable efforts that will profit the organization, and a strong desire to remain a 

7

member of the organization. More recently, Meyer and Allen (1991) attempted to isolate the three 

fundamentals of organizational commitment: "(1) affective commitment refers to the employee's emotional 

attachment to, identification with, and involvement in the organization; (2) continuance commitment refers to 

an awareness of the costs associated with leaving the organization; (3) normative commitment reflects a 

feeling of obligation to continue employment". The authors based their work on the concepts proposed in 

Kelman's studies (1958) in social psychology: compliance (or exchange), identification, and internalization. 

This three components structure of commitment has been applied to marketing exchanges: continuance, 

affective and normative commitment (Gundlach, Achrol and Mentzer, 1995; Gruen, Summers and Acito, 

2000). Nevertheless most of conceptualizations and measures of commitment to the brand focus on the 

affective component in accordance with Mowday, Steers, and Porter approach (1979). In consumer behavior 

research, this affective or attitudinal commitment refers more to a true loyalty built on a positive evaluation 

of the brand than to a false loyalty based solely on social norms or switching costs (Dick and Basu, 1994; 

Oliver, 1997). In this way, commitment to the relationship is globally defined as an individual's desire to 

maintain a strong and lasting relationship with a given brand. It goes beyond the mere framework of a 

favorable opinion of the brand (attitude) and is directly related to loyalty behavior (Jacoby and Chestnut, 

1978). Commitment is reinforced in the course of successive consumption experiences (quality, value, 

satisfaction) and relies directly on trust in the brand (Morgan and Hunt, 1994; Garbarino and Johnson, 1999). 

The five concepts described above are thus based on different psychological processes related to cognitions 

(concrete / abstract), processes (evaluation / identification), and time frames (past / present / future). Each 

one contributes specific information to the brand - consumer relationship: quality pertains to evaluating the 

excellence or superiority of the product (goods and/or services) as compared to competitors; perceived value 

includes the idea of the sacrifices made (time, effort, money) to buy and consume the product; consumer 

satisfaction is a global state derived from cumulative consumption experiences, and is both cognitive 

(evaluation) and affective (feelings); brand trust encompasses the idea of risk in all future exchanges, and 

refers in part to moral considerations (benevolence, integrity, keeping promises); consumer commitment in a 

lasting relationship with the brand is the result of an instrumental calculation (means-ends tradeoff), an 

identification (congruence of identities), and an internalization of social norms. 

8

From product perceived quality toward commitment to the brand 

In studies on the brand-consumer relationship, it is not always easy to establish and contrast the domain of 

the brand from that of the branded product. Nevertheless, the first two concepts (quality and value of the 

good or service) seem to be strongly rooted in perceptions of the product itself, whereas the last two (trust 

and commitment to the brand) appear to clearly encompass the perception of the brand. On the continuum 

from product to brand, consumer satisfaction seems to be a point of equilibrium between the realities of the 

brand and those of the product. This general framework thus accounts for the weight of both the products and 

the brands in the development of long-term relationships. In our view, an overall understanding of the 

consumer's relationship to the brand cannot be achieved without analyzing concrete evaluations of products. 

While no general model has actually been proposed and tested, the relational literature implicitly articulates 

these concepts in terms of the following causal chain: perceived quality → perceived value → cumulative 

satisfaction → trust → commitment (Figure 1), which will hereafter be called the "relational chain". Each 

successive concept is fed by the judgments immediately "upstream" from it in the causal chain, and 

incorporates additional considerations: product quality is the point of departure, and commitment to the 

brand is the ultimate stage at which the brand has firm roots (Dwyer, Schurr, & Oh, 1987; Berry, 2000). The 

four causal relationships inherent in this chain (quality → value, value → satisfaction, satisfaction → trust 

and trust → commitment) have been empirically validated, and form a point of consensus in the marketing 

research. However, most models seem to focus on one portion of the causal chain, i.e., either on consumer 

satisfaction at the beginning, or on the links between trust and commitment at the end. To our knowledge, no 

empirical application has simultaneously integrated all five concepts. 

Figure 1. Logical chain of relationship marketing 

Perceived 

Quality 

Global 

Value 

Cumulative 

Satisfaction 

Trust 

Commitment 

Concerning the upstream portion of the chain, Zeithaml (1988), Parasuraman, Zeithaml, and Berry (1988), 

who studied means-ends chains, empirically validated the "perceived quality → value" link, emphasizing that 

9

quality is one of the major antecedents of value. Oliver (1999) and Slater and Narver (2000) insisted on the 

role of perceived quality, among other factors, as a source of value enhancement (understood as a cost- 

benefit ratio) and hence greater satisfaction. Cronin and Taylor (1992) showed that it is perceived quality that 

influences satisfaction and not the reverse, contrary to the formulations of Bitner (1990) or Bolton and Drew 

(1991). Other authors have also justified the "perceived quality → satisfaction" link (Fornell, 1992; 

Anderson, Fornell, & Lehman, 1994; Anderson & Sullivan, 1993) and emphasized the role of perceived 

quality in creating a feeling of overall satisfaction. Zeithaml, Berry, and Parasuraman (1996) argued for the 

causal link "perceived quality → satisfaction → loyalty" and empirically demonstrated the "perceived quality 

→ loyalty" link. In their ACSI model, Fornell, Johnson, Anderson, Cha, and Bryant (1996) validated -- but at 

an aggregated level -- the causal chain "perceived quality → value → cumulative satisfaction → loyalty". 

Finally, Oliver (1999) proposed that value be regarded as an antecedent of satisfaction when seen as the 

overall cost-to-benefit ratio (as in this study). Note that the author also suggested that a "broader" view of 

value could make satisfaction a source of perceived value, or value and satisfaction, both functions of 

perceived quality, could be "situated at the same level" rather than linked by a causal relation. Concerning 

the downstream portion of the chain, Morgan and Hunt (1994) demonstrated the "trust → commitment" 

causality in an industrial setting. They emphasized the role of performance, value, and satisfaction as 

antecedents of trust, placing them under the heading "benefits of the relationship". Certain authors have 

nevertheless suggested inverting the causality between satisfaction and trust, using the structural model 

validated by Morgan and Hunt (1994). Accordingly, Singh and Sirdeshmukh (2000) situated trust (in the 

"cumulative" sense) as a prerequisite to "satisfaction with the transaction". These authors, who articulated the 

cumulative (acquired trust) and transactional (satisfaction with the transaction) perspectives, identified the 

cumulative relationship (in this case, trust) as an influential factor in evaluating a particular experience 

(transactional quality and satisfaction), which logically led them to reverse the causality. In sum, perceived 

quality is most often considered as the foundation of a brand - consumer relationship that is manifested 

downstream by the creation of a deep commitment to the brand. Two links have nevertheless given rise to 

contradictory conceptualizations ; value and satisfaction on the one hand, and satisfaction and trust on the 

other. This contradiction can be explained in terms of the perspective adopted by the authors: transactional, 

10

cumulative, or a combination of the two. Indeed, the cumulative status of the relationship (trust, cumulative 

satisfaction) at any given moment influences perceived performance during a particular experience, as well 

as any evaluation that follows from that relationship (quality, value, satisfaction). The present study takes a 

solely cumulative approach. 

AN EMPIRICAL STUDY ON RETAIL BANKING 

This empirical study concerns the consumer relationship to a large bank. Due to its complexity and the 

associated implications and perceived risks, the banking services sector is an interesting field for studying the 

brand-consumer relationship (Berry, 1983, 1995). In addition, as Berry (2000) emphasized, "the company 

becomes the brand" in the service sector. Banks today still stand to lose business if they are not able to 

counter their competitors' offers of products (often hard to differentiate) and to cope with critical incidents 

(Parasuraman & Grewal, 2000). From this perspective, the development of a reciprocal relationship of trust 

and commitment between the bank and its customers becomes a prime objective. 

This study pertains more to the bank as an institution than to the customer's relationship to a customer 

relations manager or a particular product. Two surveys were conducted, the second one serving as a cross- 

validation of the first. For the first survey (Phase 1) run in the spring of 2000, 30,000 questionnaires were 

sent by mail to a representative sample of 18- to 75-years old customers of a regional branch of the bank. 

The sample was randomly drawn from the general client base. A total of 2,150 completed questionnaires 

were returned, making for a response rate of 7.16%. The second phase (Phase 2) was run under similar 

conditions in the Spring of 2001. Of the 30,000 questionnaires sent out, 2,154 were returned (response rate 

7.2%). Among the respondents, there were 528 individuals who had participated in the first phase. The 

second sample was used for the cross-validation. 

Exploratory Factor Analysis 

Since the questionnaire was to be distributed by mail, by the bank itself, to its own customers, it needed to be 

reasonably short, with a limited number of items measuring each concept. The rating scales used were not 

developed by the authors but were borrowed or adapted from the literature (Appendix A1); all had been 

validated previously. Because our approach was relational, the concepts were defined and operationalized 

according to a "cumulative" logic. Our formulation of the items was not based on a "particular consumption 

11

experience" (from a transactional perspective) but on the cumulative, overall relationship with the bank as a 

whole, from a perspective that spans different time frames. In order to maintain the rigor of the validity test, 

the items were mixed up in the questionnaire and were not presented "construct by construct", given that the 

latter practice tends to have a direct impact on the internal consistency of the ratings. As a first step, 

exploratory factor analyses were conducted on the first sample (phase 1), to refine the seven constructs 

(technical and functional quality, value, satisfaction, credibility and benevolence, commitment) taken in 

isolation, then two by two, three by three, etc. These analyses led us to delete one functional-quality item that 

showed too little correlation with the other two (funcqual1). In accordance with Grönroos's (1983) theoretical 

framework, perceived quality can be broken down into "technical quality" and "functional quality". By 

contrast, trust consists of a single dimension; this runs contrary to the results found by Ganesan and Hess 

(1997), who distinguished credibility and benevolence. The global exploratory factor analysis of the six 

constructs (Table 1) indicated that, despite their number, their conceptual proximity, the use of a single 

method (Likert scale), and the mixing up of the items in the questionnaire, the constructs demonstrated an 

acceptable degree of convergent and discriminant validity. All correlations between the items and the factor 

they are assumed to measure were high; inversely, all correlations between these items and the other factors 

were less than 0.25. Cronbach's alpha coefficients calculated on this basis are shown in Table 1. They range 

between .73 and .90, which can be considered satisfactory. 

Table 1. Principal component analysis on all items and Cronbach’s Coefficients (Phase 1, 2000) , oblique 

rotation, correlations below .25 are not shown). 

Item factor1 factor2 factor3 factor4 factor5 factor6 

Cronbach’s alpha 0.77 0.75 0.90 0.89 0.73 0.82 

techqual1 .93 

techqual2 .73 

funcqual1 .84 

funcqual2 .90 

funcqual3 

val1 -.98 

val2 -.81 

satis1 -.98 

satis2 -.60 

satis3 -.61 

trust1 -.87 



12


commit1 .59 

commit2 .51 

commit3 .92 

Confirmatory Factor Analysis 

We tested the unidimensionality (in the sense defined by Anderson, Gerbing, and Hunter, 1987) of each 

construct separately, then two by two, three by three, etc., and finally for all six constructs taken together. 

The confirmatory factor analysis were performed by Lisrel 8 software. The estimation was performed using 

the maximum likelihood method because of its robustness for large sample sizes (n = 2,150 and n = 2,154 for 

Phases 1 and 2, respectively). The results are presented in Table 2. For Phase 1, the values of the goodness-of 

fit indexes, RMSEA, AGFI, and RMR, are quite satisfactory, as were the principal indexes of the goodness 

of fit with a saturated model (non-normed fit index, comparative fit index). We can therefore accept the 

overall unidimensionality of the six constructs. Note, however, that we had to delete one item from the 

commitment construct (attach3). The Phase 2 estimation, conducted for the purposes of cross-validation, was 

computed on exactly the same items. The result was confirmed and the fit was slightly better. Having 

demonstrated the unidimensionality of the constructs, we then went on to determine their psychometric 

properties. 

Convergent validity. The six constructs exhibited good convergent validity. The lambda coefficients 

(equivalent to the notion of saturation in factor analysis) were all above 0.70 and are highly significant. The 

reliability coefficients of the composite factors ranged between 0.72 and 0.90 (percentage of the variance 

shared by all items measuring a given factor). The mean variance indicators, which assess the portion of true 

variance extracted from the questions measuring a construct with respect to the margin of error (Fornell and 

Larker, 1981), were all above 0.57 (Table 3), which is still within the satisfactory range. Very similar results 

were found in Phase 2, the cross-validation phase. We can see considerable stability in the reliability 

coefficients and variance coefficients obtained (Table 3). 

Discriminant validity. Here, the results are not as good due to the higher correlations between the constructs. 

Comparing the mean variance indicators obtained (shared variance indexes) with the squares of the inter- 

construct correlations (see Table 4), we can see that technical quality, global value, satisfaction, trust, and 

13

commitment share about the same amount of variance with each other as with their own measures 

(sometimes a little more, sometimes a little less). This lack of discriminant validity was confirmed by the 

cross-validation (Phase 2), where we observed very similar correlation levels (Table 4). These results, 

observed for a large sample of "real" customers, allay doubts as to the discriminant validity of the constructs 

and to the magnitude of their respective contributions to explaining consumer behavior, since they share 

essentially the same variance. To put this strict interpretation into perspective, note that the high correlations 

can be partly attributed to the fact that the items were measured via a single method (Likert scale), which 

artificially increases correlations between constructs, through a common "method" effect. 

Table2. Fit indicators of the measurement model (confirmatory factor analysis) 

Phase 1 (2000) Phase 2 (2001) 

Chi-square,75 df (prob) 389 (.0) 343 (.0) 

RMSEA (prob) .044 (.99) .041 (1) 

standardized RMR .021 .019 

AGFI .96 .97 

CFI .98 .99 

RFI .94 .98 

Table 3. Psychometric properties obtained in the confirmatory factor analysis 

Phase 1 (2000) Phase 2 (2001) 

No. of Coefficient of Coefficient of Coefficient of Coefficient of 

items reliability () shared variance reliability shared variance 

Functional quality 2 .75 .60 .73 .58 

Technical quality 2 .72 .57 .76 .62 

Global value 2 .81 .68 .83 .72 

Satisfaction 3 .88 .71 .89 .73 

Trust 4 .90 .68 .90 .69 

Commitment 2 .74 .59 .73 .57 

Table 4. Correlations between constructs (confirmatory factor analysis), Phase 1 (2000) and Phase 2 (2001) 

Technical Functional Global value Satisfaction Trust 

quality quality 

Technical quality 1 

Functional quality .52 (.56) 1 

Global value .77 (.79) .58 (.59) 1 . 

Satisfaction .75 (.78) .66 (.65) .83 (.82) 1 

Trust .70 (.75) .66 (.63) .72 (.73) .82 (.83) 1 

Commitment .75 (.81) .67 (.69) .81 (.83) .81 (.86) .84 (.86) 

Thus, we found contrasted results that fit well enough to allow us to validate the unidimensionality of the 

constructs but with high correlations, suggesting a lack of discriminant validity. During both research phases, 

14

we observed that the least correlated constructs were the two dimensions of product quality -- technical and 

functional (r = 0.52, see Table 4). The other constructs (value, satisfaction, trust, and commitment) were 

strongly and approximately equally correlated (between 0.72 and 0.84 for Phase 1 and between 0.73 and 0.86 

for Phase 2), and as we have seen, lack discriminant validity. This observed lack of discriminant validity 

answers our initial question. At the theoretical level, each concept has its own meaning: the excellence of 

products and services (perceived quality), the cost-benefit tradeoff (perceived value), the fulfillment of 

individual goals through consumption experiences (cumulative satisfaction), the credibility and perceived 

benevolence of the brand (trust), and the desire to maintain a strong and lasting relationship with the brand 

(commitment). On the other hand, at the empirical level, a simultaneous analysis of all of these concepts 

indicated strong correlations between them, casting doubt on the usefulness of having separate concepts in 

our attempts to understand the consumer relationship to the brand. As soon as the questionnaire moved away 

from concrete evaluations of products in order to assess abstract constructs related to brands, the respondents 

seemed to have more difficulty discriminating the concepts (satisfaction, trust, commitment). While these 

theoretical developments are relevant, applying them to surveys may seem at least partly superfluous in 

studying the brand-consumer relationship and predicting its effects on consumer behavior. 

Testing the Relational Chain "Quality → Value → Satisfaction → Trust → Commitment" 

Our study of the literature shows that few authors have devised and empirically tested an overarching model 

of the relationship to the brand. Most researchers prefer to visualize the principal constructs in pairs (quality 

and value, trust and commitment, etc.) Nevertheless, the literature does propose a theoretical model in the 

form of a causal chain: quality → value → satisfaction → trust → commitment, as we discussed in Section 1 

(see figure 1 below). In order to validate this chain, we began by testing a completely recursive model 

(depicted as solid lines in Figure 2) which represents the (strong) hypothesis that the effect of each concept in 

the upstream portion of the chain is completely mediated by the concepts in the downstream portion. 

Although this model is very restrictive, it is faithful to the idea that the fundamental goal of relationship 

marketing is to build a lasting relationship through the accumulation of positive experiences, based on a 

quality management policy that leads to lasting customer commitment to the brand (Dwyer, Schurr, and Oh, 

1987). The Phase 1 assessment showed that the model, in its most restrictive form, is not satisfactory. We 

15

had to make it more flexible by allowing both facets of perceived quality to have direct effects on 

satisfaction, trust, and commitment (dotted lines in Figure 2). This made for a satisfactory estimation, 

including in Phase 2 (Table 5). We clearly observed a "relational chain" structure with significant direct 

effects of both quality components (technical and functional) on value, and effects of value on satisfaction, of 

satisfaction on trust, and finally, of trust on commitment (Figure 2). Indeed, all these links are positive, 

strong, and nearly equal in intensity (close to 0.5). The same results were obtained in Phase 2 (values shown 

in parentheses in Figure 2). 

Figure 2. Structure of the model of the relationship marketing chain (standardized coefficients for Phase 1 

(Phase 2 in parentheses)) 

Technical 


Functional 


.69 (.71) 

.23 ( 19) 

.27 (.36) 

Perceived 

Value 

.47 (.39) .54 (.56) 

Satisfactio 

n 

.24 (.21) .20 (.13) 

.19 (.24) .38 (.43) 

Trust 

.45 (.42) 

.19 (.19) 

Table 5. Test of model : Relational chain: quality -> value -> satisfaction -> trust -> commitment 

Phase 1 (2000) Phase 2 (2001) 

Chi-square, 78 df (prob) 447 (.00) 389 (.00) 

RMSEA (prob) .048 (.79) .044 (.99) 

standardized RMR .023 .021 

AGFI .96 .96 

CFI .98 .99 

RFI .98 .98 

Commitment 

However, this causal chain does not in itself account for the full impact of perceived quality on the relational 

process. Technical and functional quality still have significant direct effects on satisfaction, trust, and 

commitment. These direct effects characterize the impact of the perceived quality of products and services -- 

16

outside of the consumption experience -- on the relationship between the customer and the institution, and 

they confirm the fundamental role played by quality in building a lasting relationship with the customer. A 

comparison of the magnitude of these direct effects showed that technical quality (bank products) plays a 

more important role than functional quality (bank services), particularly with the former's strong impact on 

perceived value. Note also that the direct effect of technical quality on commitment was almost as great as 

that of trust. Added to the indirect effects of technical quality via the components of the relational chain, this 

direct effect gave us a total effect of technical quality on commitment of 0.61 versus 0.36, 0.11, 0.24, and 

0.47 for functional quality, value, satisfaction, and trust, respectively (keeping in mind that the parameter 

values are partially affected by the theoretical structure of the model tested). Thus, among the concepts 

studied here, technical quality had the strongest effect on commitment to the brand, a concept shown in the 

literature to be strategic in building a lasting relationship with customers. The same result having been 

observed in Phase 2, we can conclude that technical quality remains a fundamental indicator to be monitored, 

for any decline in technical quality has potentially major consequences at the downstream end of the chain. 

The model of the relational chain -- which is the most common in the existing literature -- emphasizes the 

fundamental role played by perceived quality at the managerial level, in so far as it is the basis for 

developing and maintaining the relationship with the customer. Furthermore, because the causal structure is 

in the form of a chain ending with commitment -- the final stage in the development of the relationship 

according to Berry (2000) -- all the links in the chain must be reinforced: quality, value, satisfaction, and 

trust. Like all chains, it may become more fragile if any one of its links weakens, putting the entire relational 

structure in jeopardy. An example might be the case of a customer who left his bank after years of 

satisfactory dealings leading to a strong commitment, because he perceived more value in what a competitor 

had to offer. 

CONCLUSION 

Working in the field of relationship marketing, we tested the validity of five major concepts characterizing 

the customer's relationship to the brand: perceived quality (both technical and functional), global value, 

satisfaction, trust, and commitment. While we were able to show their unidimensionality and their 

convergent validity, their strong correlations (ranging from 0.70 to 0.80) cast doubt on their discriminant 

17

validity. This raised a question about the respective contributions of concepts which share more than 50% of 

their variance, whether in characterizing the customer-brand relationship at a given moment or in predicting 

its impact on consumer behavior, particularly on loyalty (repurchase, price tolerance, resistance to change, 

positive word of mouth). An analysis of the relationships between these concepts showed that, if we admit 

that causal structure, the sequence "quality → value → satisfaction → trust → commitment" is the one that 

fits our data the best. Nonetheless, this causal structure merits further development at the theoretical level. In 

particular, we suggest that the causal structure of the chain may vary according to whether it is viewed from 

a transactional or cumulative perspective. 

This study suffers from several limitations and brings up a number of methodological questions for future 

research in relationship marketing. 1) We studied the brand-consumer relationship at a single moment, even 

though it is more of an incremental process, as Weitz and Jap (1995) already emphasized in their critique of 

the studies by Morgan and Hunt (1994) or Fournier (1998). 2) At the time of our surveys, it is difficult to 

guarantee that the object evaluated (even though it was clearly stated in the questionnaire) was the same for 

all respondents: Was it the institution, the brand, a particular product, a customer service manager? 3) The 

use of a single method could have artificially inflated the correlations, thus undervaluing the discriminant 

validity. Using the same method (Likert scales) to measure the same object (the relationship to the brand) at 

the same moment (one questionnaire) can create halo effects. 4) Since the response rate to the questionnaire 

was only 7.7%, our sample certainly suffered from representativity problems, even though it maintained the 

division of the population according to the major socio-demographic criteria. 5) Since banks mainly provide 

a "utilitarian" service, it would be interesting to see how the results evolve for more experiential products or 

services in fields like consumer food products, sports, leisure activities, or artistic pursuits. 

18

REFERENCES 

Anderson Eugene W., Fornell Claes, and Lehman Donald R. Customer satisfaction, market share and 

profitability: findings from Sweden. Journal of Marketing 1994; 58 (July): 53-66. 

Anderson Eugene.W. and Sullivan Mary W. The antecedents and consequences of customer satisfaction for 

firms. Marketing Science 1993; 12 (Spring): 125-143. 

Anderson James C., Gerbing David and Hunter John E. On the assessment of unidimensional measurement: 

internal and external consistency and overall consistency criteria. Journal of Marketing Research 1987; 24 

(November): 432-437. 

Becker Gary S. The Economic Approach to Human Behavior. Chicago: University of Chicago Press, 1976. 

Berry Leonard L. Relationship marketing. In: G.L. Shostack and G. Upah, editors. Perspectives On Services 

Marketing. Emerging, Chicago: AMA, 1983. 

Berry Leonard L. Relationship marketing of services - growing interest, emerging perspectives. Journal of 

the Academy of Marketing Science 1995; 23 (Fall): 236-245. 

Berry Leonard L. Cultivating service brand equity. Journal of the Academy of Marketing Science 2000; 28 

(Winter): 128-137. 

Bitner Mary J. Evaluating service encounters: the effects of physical surroundings and employee responses. 

Journal of Marketing 1990; 54 (April): 69-82. 

Bolton Ruth N. and. Drew James H. A multistage model of customers' assessment of service quality and 

value. Journal of Consumer Research 1991; 17 (March): 375-384. 

Boulding William, Kalra Ajay, Staelin Richard, and Zeithaml Valarie A. A dynamic process model of 

service quality: from expectations to behavioral intentions. Journal of Marketing Research 1993; 30 

(February): 7-27. 

Brady Michael K. and Cronin Joseph J. Some new thoughts on conceptualizing perceived service quality: a 

hierarchical approach. Journal of Marketing 2001; 65 (July):34-49. 

Cronin Joseph J. and Taylor Steven A. Measuring service quality: a reexamination and extension. Journal of 

Marketing 1992; 56 (July): 55-68. 

19

Dick Alan S. and Basu Kunal. Customer loyalty: toward an integrated conceptual framework. Journal of the 

Academy of Marketing Science 1994; 22 (Spring): 99-113. 

Dwyer F. Robert, Schurr Paul H. and Oh Sejo. Developing buyer-seller relationships. Journal of Marketing 

1987; 51 (April): 11-27. 

Fornell Claes. A national satisfaction barometer: the swedish experience. Journal of Marketing 1992; 56 

(January): 1-21. 

Fornell Claes, Johnson Michael D., Anderson Eugene W., Cha Jaesung and Bryant Barbara Everitt. The 

american customer satisfaction index: nature, purpose, and findings. Journal of Marketing 1996; 60 

(October): 7-18. 

Fornell Claes and Larcker David F. Structural equation models with unobservable variables and 

measurement error: algebra and statistics. Journal of Marketing Research 1981; 18 (August): 382-388. 

Fournier Susan. Consumers and their brands: developing relationship theory in consumer research. Journal of 

Consumer Research 1998; 24 (Mars): 343-373. 

Fournier Susan and Mick David Glen. Rediscovering satisfaction. Journal of Marketing 1999; 63 (October): 

5-23. 

Ganesan Shankar. Determinants of long-term orientation in buyer-seller relationships. Journal of Marketing 

1994; 58 (April): 1-19. 

Ganesan Shankar and Hess Ron. Dimensions and levels of trust: implications for commitment to a 

relationship. Marketing Letters 1997; 8 (October): 439-448. 

Garbarino Ellen and Johnson Mark S. The different roles of satisfaction, trust and commitment in customer 

relationships. Journal of Marketing 1999; 63 (April): 70-87. 

Grönroos Christian. Strategic Management and Marketing in the Service Sector. Cambridge, MA: Marketing 

Science Institute, 1983. 

Gruen Thomas W., Summers John O. and Frank Acito. Relationship marketing activities, commitment, and 

membership behavior in professional associations, Journal of Marketing 2000; 64 (July): 34-49. 

Gundlach Gregory T., Achrol Ravi S. and Mentzer John T. The structure of commitment in exchange. 

Journal of Marketing 1995; 59 (January): 78-92. 

20

Holbrook Morris B. Consumer Value: A Framework for Analysis and Research. Routledge Interpretative 

Market Research Series, 1999. 

Iacobucci Dawn, Grayson Kent A. and Ostrom Amy L. The calculus of service quality and customer 

satisfaction: theoretical and empirical differentiation and integration. Advances in Services Marketing and 

Management 1994; 3: 1-67. 

Jacoby J. and Chestnut R. Brand Loyalty: Measurement and Management. New York: John Wiley, 1978. 

Johnson Michael D., Anderson Eugene W. and Fornell Claes. Rational and adaptive performance 

expectations in a customer satisfaction framework. Journal of Consumer Research 1995; 21 (March): 695- 

707. 

Kelman H.C. Compliance, identification, and internalization: Three Processes of Attitude Change. Journal of 

Conflict Resolution 1958; 2: 51-60. 

Lai Albert Wenben. Consumer values, product benefits and customer value: a consumption behavior 

approach. In: F.R. Kardes and M. Sujan, editors. Advances In Consumer Research, vol. 22., Provo, UT: 

Association for Consumer Research, 1995. pp. 381-388. 

Meyer John P. and Allen Natalie J. A three-component conceptualization of organizational commitment. 

Human Resource Management Review 1991; 1 (Spring): 61-89. 

Morgan Robert M. and Hunt Shelby D. The commitment-trust theory of relationship marketing. Journal of 

Marketing 1994; 58 (July): 20-38. 

Mowday Richard T., Steers Richard M. and Porter Lyman W. The measurement of organizational 

commitment. Journal of Vocational Behavior 1979; 14: 224-247. 

Oliver Richard L. A conceptual model of service quality and service satisfaction; compatible goals, different 

concepts. Advances In Services Marketing And Management 1993; 2: 65-85. 

Oliver Richard L. Satisfaction: A Behavioral Perspective on the Consumer. New York: MacGraw Hill, 1997. 

Oliver Richard L. Value as excellence in the consumption experience. In: Holbrook Morris B., editor. 

Consumer Value: a Framework for Analysis and Research. Routledge, 1999, pp. 43-62. 

21

Parasuraman A. and Grewal Dhruv. Serving customers and consumers effectively in the twenty-first century: 

a conceptual framework and overview. Journal of the Academy of Marketing Science 2000; 28 (Winter): 9- 

16. 

Parasuraman A., Zeithaml Valarie A. and Berry Leonard L. Servqual: A multiple-item scale for measuring 

consumer perceptions of service quality. Journal of Retailing 1988; 64 (Spring): 12-40. 

Singh Jagdip and Sirdesmukh Deepak. Agency and trust mechanisms in consumer satisfaction and loyalty 

judgments. Journal of the Academy of Marketing Science 2000; 28 (Winter): 150-167. 

Slater Stanley T. and Narver John C. Intelligence generation and superior value. Journal of the Academy of 

Marketing Science 2000; 28 (Winter): 120-127. 

Zeithaml Valarie A. Consumer perceptions of price, quality, and value: a means-end model and synthesis of 

evidence. Journal of Marketing 1988; 52 (July): 2-22. 

Zeithaml Valarie A., Berry Leonard L. and Parasuraman A. The behavioral consequences of service quality. 

Journal of Marketing 1996; 60 (April): 31-46. 

22

APPENDIX 1. DESCRIPTION OF THE ITEMS 

Perceived quality. We adopted the approach taken by Grönroos (1983), who distinguished two dimensions: 

(1) core-service performance -- here banking products (savings accounts, stocks, homeowner's loans, and life 

insurance, 2 items) -- which corresponds to the notion of technical quality, and (2) quality as manifested in 

the bank's physical service provisions (arrangement of lobby, cleanliness, and general welcome in the bank's 

branches, 3 items), which corresponds to the notion of functional quality. 

Perceived global value. This dimension was measured in a cumulative perspective, as the ratio of received 

benefits to agreed sacrifices in the course of the customer's dealings with the service provider (Zeithaml, 

1988; 2 items). 

Cumulative satisfaction. Derived from Oliver's (1997) rating scale, cumulative satisfaction is affected by the 

possible regret felt from having established a commercial relationship with the bank (3 items). 

Trust. Trust was measured through its two main facets, based on the work by Ganesan and Hess (1997): the 

service provider's credibility (2 items) and benevolence (2 items). 

Commitment. Commitment refer to the “relative intensity of identification and involvement” with the bank 

(Mowday, Steers, and Porter, 1979). It was operationalized as a whole with the use of three items that mean 

the share of values and goals of the bank and a sense of membership. 

23

Construct Item How construct is approached in questionnaire 

Technical qualitytechqual1 

If I have to take out a loan to buy a house or apartment, Bank XX will 

(Grönroos) 

offer me a mortgage at an advantageous rate 

techqual2 If one day I have to buy some stocks or bonds, I think I would go to 

Bank XX to accomplish this operation 

Functional qualityfuncqual1 

I always feel welcome in the branch offices of Bank XX 

(Grönroos) 

funcqual2 The branch offices of Bank XX are particularly well kept 

funcqual3 The branch offices of Bank XX are well designed and furnished 

Global value (Zeithaml) val1 The products and services of Bank XX are generally offered at a good 

quality-to-price ratio 

val2 The services provided by Bank XX are good for the price they charge 

Cumulative satisfactionsatis1 

Since I've been a customer of Bank XX, I've had no complaints about 

(Oliver) 

the management of my checking account 

satis2 I really made the right decision in choosing Bank XX as my bank. 

satis3 I'm satisfied with the services provided by Bank XX 

Trust (Ganesan & Hess) trust1 My account advisor at Bank XX performs his/her job with real 

integrity 

Credibility trust2 I know I can count on the promises my financial advisor makes me 

Benevolence trust3 My counselor at Bank XX takes my interests into account 

trust4 My counselor at Bank XX is always ready to help me in case of a 

problem 

Commitment (Mowday, commit1 At Bank XX, I feel like I'm more than just a customer 

Steers and Porter, 1979) 

commit2 I share the values stated by Bank XX 

commit3 Bank XX is clearly different from other banks 

Note.: The scales have been adapted to a European context and to a retail banking market and translated into 

a European language (and translated again in English in this appendix) 

24

An Empirical Examination of Retail Convenience for In-Store and Online Shoppers 


Michelle Bednarz, Mississippi State University 

Abstract: In this study, we develop and assess multidimensional scales of retail convenience for 

in-store and online shoppers. Principal components analysis and confirmatory factor analysis reveal 

five dimensions of retail convenience for in-store shoppers, and four dimensions of retail 

convenience for online shoppers. Analysis of variance results indicate that online shoppers benefit 

more from access convenience and search convenience than in-store shoppers. Implications for 

retailers and areas for future research are also addressed. 

Key Words: convenience, consumer decision-making, online shopping, scale development 

Introduction 

The concept of convenience has been an important topic in marketing for a time period 

spanning approximately 80 years (Copeland 1923; Gardner 1945; Kelley 1958; Brown 1989, 1990; 

Seiders, Berry, and Gresham 2000; Berry, Seiders, and Grewal 2002). However, much of this 

research focuses on convenience goods, rather than convenience as a benefit realized by the 

consumer. Because today’s consumer is more time-starved than ever, it is appropriate to more 

carefully consider convenience as a concept of utmost importance. 

Foote (1963) characterized the consumer of the future (in the year 2000) as one whose 

primary constraints are no longer money, but time and learning. His prediction is correct. Today’s 

consumers consider the resource of time as valuable as money (if not more valuable). Bhatnagar, 

Misra, and Rao (2000, p.98) confirm this by stating that “the consumer maximizes his utility subject 

to not only income constrains but also time constraints.” 

As time becomes a more valuable consumer resource, convenience also provides added 

value. Michael J. Sansolo, senior vice-president at the Food Marketing Institute, illustrates the 

degree of importance consumers place on convenience by stating, “Give shoppers a choice between 

lower prices or more convenience and convenience will win every time” (Forster 2002, p. 120). 

Online shopping is particularly well suited to offer convenience to time-starved consumers 

(Kaufman-Scarborough and Lindquist 2002). In an otherwise lackluster economy, the strong 

25

growth rates of companies such as eBay, Land’s End, and Amazon have occurred because they 

provide customers with what they want – its all about convenience (Green 2002). 

When examining the literature related to the convenience construct, many gaps exist, 

particularly in a retail context. While it is assumed that consumers often seek convenience in 

shopping situations, we know little about the individual dimensions of retail convenience. Seiders, 

Berry, and Gresham (2000) propose four dimensions of convenience that are particularly relevant 

for retailers: access, search, possession, and transaction convenience (see Table 1). An empirical 

investigation of how these convenience dimensions provide a benefit to consumers is warranted. 

Specifically, do in-store shoppers and online shoppers seek and obtain different types of 

convenience? This study considers that different structures of retail convenience may exist, 

depending on the shopping format chosen by the consumer. The argument developed here is that 

different shopping formats deliver different types of convenience, and a consumer’s preference for a 

specific type of convenience determines his or her decision to choose a particular shopping format – 

either online or at a traditional store. 

This study is important for several different reasons. First, it follows the logical progression 

of research in this area. Retailers may now better understand the existence of different convenience 

dimensions, but how can they use this new knowledge to better satisfy customers? As Berry (2001, 

p. 136) states, “Superior retailers understand that people’s most precious commodity in the modern 

world is time and do everything they can to save as much of it as possible for their customers.” By 

developing a better understanding of the different dimensions of retail convenience and by 

understanding which dimensions of convenience are most important to customers, retailers will be 

able to better understand how to meet customers’ needs, thus improving customer satisfaction. If 

properly implemented, retail convenience could become a source of sustainable competitive 

advantage for retailers. 

26

This paper is organized in the following manner: A review of the literature relevant to this 

study is presented. Scale development procedures are followed in order to more carefully examine 

the specific dimensions of retail convenience for both in-store and online shoppers. A different 

model for each type of shopping format is hypothesized and tested using structural equation 

modeling techniques. Results are given, followed by a discussion that addresses the implications of 

the findings for retailers with traditional stores and/or an online presence. Limitations and 

directions for future research are also provided. 

Literature Review 

The concept of convenience first appeared in the marketing literature with Copeland’s (1923) 

classification of goods. Copeland suggests that by classifying goods according to his tripartite 

structure (convenience, shopping, or specialty goods), marketers can determine the type of store in 

which the product should appear and can determine the appropriate concentration of distribution. 

Convenience goods are those lower-priced goods which consumers are familiar with and which are 

purchased from easily accessible outlets. Gardner (1945, p. 275) provides a further description: 

Convenience goods are articles of daily purchase…which are insignificant in value 

or are needed for immediate use. These goods are, to a considerable extent, bought 

at the most convenient place without a comparison of values…. 

As marketers continued to develop other product classification systems, convenience goods 

remained an essential staple, re-appearing in several other schemata (Bucklin 1963; Kaish 1967; 

Holbrook and Howard 1977; Enis and Roering 1980; Murphy and Enis 1986). Thus, the initial use 

of the word “convenience” in the marketing literature was as an adjective describing a class of 

consumer goods. Many researchers have conducted studies covering the vast domain of 

convenience goods, including convenience foods (e.g., frozen dinners, ready-to-eat cold cereal, fast- 

food restaurants) (Crist 1960; Anderson 1972; Reilly 1982; Darian and Cohen 1995), convenience 

time-saving durables (e.g., dishwashers, microwave ovens, washers and dryers) (Anderson 1972; 

27

Reilly 1982), and time-saving services (e.g., child care, house cleaning services, lawn care services) 

(Brown 1990). 

Over time, the use of the word “convenience” has changed/evolved from a descriptor of 

products into its own unique concept. The more complete definitions of convenience contain one 

common element - they all view convenience as any attribute that reduces the non-monetary costs 

associated with a product (Kelley 1958; Kotler and Zaltman 1971; Etgar 1978). For purposes of 

this study, retail convenience (at its most general level of abstraction) is defined as the consumer’s 

perceptions of time and effort costs associated with making a purchase. As the reduction of time 

and effort costs is a motivating factor for consumers to seek convenience, these concepts are 

described more fully below. 

Time and Effort Literature 

Time is a limited, scarce, and therefore valuable consumer resource (Jacoby, Szybillo, and 

Berning 1976). As a limited, scarce, and valuable resource, consumers are always looking for ways 

to save time, be it through convenience products, convenience services, convenience durables, or 

more recently through convenience shopping. One reason why saving time has become so 

important to consumers is because time, unlike money or other resources, is fixed and therefore 

incapable of being expanded (Berry 1979; Gross 1987). According to Berry (2001, p.136), 

“Superior retailers understand that people’s most precious commodity in the modern world is time 

and do everything they can to save as much of it as possible for their customers.” 

Like time, effort is also a valuable consumer resource which is characterized by consumers’ 

energy expenditures. Activities requiring great amounts of energy on the consumers’ part are 

considered inconvenient, while activities requiring minimal energy on the consumers’ part are 

considered convenient. As a completely different type of non-monetary cost, consumers’ effort 

expenditures not only influence perceptions of convenience (Seiders, Berry, and Gresham 2000), 

but also influence levels of satisfaction (Lovelock 1983). 

28

Human and cognitive effort has been the topic of many studies appearing in other disciplines 

including operations management, psychology, decision theory, and economics (Youngdahl and 

Kellog 1997; Bettman, Johnson, and Payne 1990). However, consumer effort expenditures have 

received significantly less attention within the marketing literature (Berry, Seiders, and Grewal 

2002), perhaps because product or service attributes that aim to save effort are perceived as being 

the same as, and as a result are lumped together with, those that aim to save time (Brown 1990). 

Thus, in a retail context, consumers view both time and effort as costs that are associated with 

obtaining the desired product and/or service. In the decision-making process, consumers must 

decide if the costs of obtaining the desired product/service are worth the benefits they will receive 

from the purchase. 

Cost-Benefit Analysis 

The driving force that motivates consumers to seek convenience involves saving both time 

and effort expenditures. According to Prest and Turvey (1965, p. 683), “Cost-benefit analysis is a 

practical way of assessing the desirability of projects” and “implies the enumeration and evaluation 

of all the relevant costs and benefits.” When deciding between several alternatives (e.g., whether to 

shop in a traditional store or in a virtual store), consumers determine the costs and benefits 

associated with each and compare those costs to the benefits. 

In the decision-making process, time and effort are significant consumer costs and are often 

considered along with money and other resources (Jacoby, Szybillo, and Berning 1976; Bhatnagar, 

Misra, and Rao 2000). Consumers choose online shopping because it facilitates the steps in the 

decision process, saving consumers both time and effort in the search, alternative evaluation, and 

purchase stages. Bhatnagar, Misra, and Rao (2000) apply cost-benefit analysis to explain why 

consumers choose to shop online: 

[Internet stores] tend to reduce the time the consumer spends on shopping (travel 

time, time spent parking, time spent traveling from the parking lot to the store, time 

spent in the checkout lines) either directly or indirectly. … The only time component 

29

emaining is the time spent browsing the Web sites (which corresponds to the time 

spent browsing the aisles in the more traditional sources of retailing). Therefore, a 

great attraction of the Internet is the convenience that it affords (p. 98). 

In other words, the benefits of shopping online (easy access to retailers) outweigh the costs (not 

obtaining the product immediately). Depending on what type of convenience is most important to 

the consumer, the cost-benefit analysis could lead to a different shopping format decision. The 

different types of convenience that consumers may choose to seek are described below. 

The Multidimensional Nature of the Convenience Construct 

Table 1 presents the various definitions and dimensions of convenience proposed by 

researchers in this area. Brown (1989) is the first to define the construct of convenience, and he 

focuses on the need for a definition that reflects the term’s multidimensional nature. The definition 

of convenience proposed by Brown (1989, 1990) contains five different dimensions: time, place, 

acquisition, use, and execution. Note that all five of the dimensions reflect the idea of saving 

consumers’ time and effort expenditures. 

-------------------------------- 

Table 1 about here 

-------------------------------- 

Berry, Seiders, and Grewal (2002) identify and define five dimensions of convenience 

applicable to the services arena: access, decision, transaction, benefit, and post-benefit. Seiders, 

Voss, Grewal, and Godfrey (2003) empirically investigate these dimensions of service convenience 

and their relationship to value perceptions and behavioral intentions. Their findings suggest that 

decision convenience has a positive effect on value perceptions; access convenience has no 

significant effect on either value perceptions or behavioral intentions; transaction convenience has a 

positive effect on behavioral intensions; and benefit convenience has a strong positive effect on 

both value perceptions and behavioral intentions. 

30

Seiders, Berry, and Gresham (2000) identify and define four dimensions of convenience that 

are specific to retailers. These four dimensions of retail convenience are essential for the purposes 

of this study. They are discussed in greater detail below. 

Access convenience. Access convenience is defined as “the speed and ease with which 

consumers can reach a retailer” (Seiders, Berry, and Gresham 2000, p. 81). This access may occur 

in person, over the phone, through a computer, or in other ways. Access convenience is an 

extremely important dimension of retail convenience, because if the consumer cannot reach the 

retailer, then all other dimensions of retail convenience are meaningless. In other words, if the 

consumer cannot reach the retailer, then they will never be given the opportunity on that particular 

shopping attempt to make a decision, to complete a transaction, or to possess the desired product 

from the retailer. 

Consumer decision-making is significantly influenced by both the speed and ease with 

which consumers can make contact with retail outlets. Traditional retailers may improve access 

convenience by operating from a location that is easy to get to, near to most consumers, and near to 

other frequently visited stores (Seiders, Berry, and Gresham 2000). Online retailers are particularly 

suited for access convenience, as it provides the opportunity for consumers to shop at home 24 hr/7 

days a week (Hofacker 2001). Morganosky and Cude (2000) found that the main reason consumers 

choose to purchase groceries online is that it eliminates travel time to and from the store. 

Although it is an important aspect of retail convenience, providing access convenience alone 

will not necessarily lead to sales. A virtual or physical store may be easy to access, but at the same 

time slow or difficult to use (Seiders, Berry, and Gresham 2000). In order to facilitate the decision- 

making process, the retailer must also provide the information necessary for the consumer to make 

the best purchase decision. 

Search convenience. After access convenience reduces the time and effort necessary to 

reach a retailer, search convenience then eases consumers through the shopping process by helping 

31

them make their purchase decision. Search convenience is “the speed and ease with which 

consumers identify and select products they wish to buy” (Seiders, Berry, and Gresham 2000, p. 

83), and includes effective interactive customer systems, store design and layout, product displays, 

store signage, and knowledgeable salespeople. 

Search convenience is extremely important for retailers operating virtual stores. One of the 

greatest consumer benefits of shopping online is the advanced search capability and the ease with 

which consumers can locate product information (Hof 2001; Shop.org 2001; Burke 2002). 

Wolfinbarger and Gilly (2001, p. 35) support this by stating that the “the online medium facilitates 

utilitarian behavior as search costs for product information are dramatically reduced.” Thus, 

shopping on the Internet facilitates search convenience. 

Possession convenience. Seiders, Berry, and Gresham (2000, p. 85) define possession 

convenience as “the speed and ease with which consumers can obtain desired products.” Included 

within the domain of possession convenience are in-stock merchandise, timely production, and 

timely delivery of merchandise. One of the main reasons why consumers choose traditional stores 

over virtual stores is the ability to actually leave the store with the desired product. Online shoppers 

must wait for their order to be processed and then wait for delivery to their home or office. For 

consumers shopping online, low possession convenience may be seen as a cost rather than a benefit. 

Transaction convenience. Transaction convenience is defined as “the speed and ease with 

which consumers can effect or amend transactions” (Seiders, Berry, and Gresham 2000, p. 86) and 

includes the checkout process as well as the return process. Hence, transaction convenience 

addresses the ease of doing business with a retailer during and/or after the act of purchase. One 

benefit of online shopping is that consumers never have to wait in line at virtual stores, thus 

enhancing transaction convenience (Wolfinbarger and Gilly 2001). Traditional stores as well as 

virtual stores with quick checkouts and easy return policies also rank high in transaction 

convenience. 

32

The four dimensions of retail convenience share a common element - they each save the 

consumer time and effort in a unique way. Whether shopping online or in a physical store, 

consumers seek these various convenience dimensions to reduce time and effort costs associated 

with consumer decision-making. The next section addresses the different models of retail 

convenience for each shopping format. 

A Model of Retail Convenience for In-Store and Online Shoppers 

Consumers often cite convenience as a major benefit of shopping online (Szymanski and Hise 

2000; Hoffman 2000; Childers, Carr, Peck, and Carson 2001; Wolfinbarger and Gilly 2001). We 

argue that a consumer’s preference for a specific type of convenience influences his or her 

preference for a particular shopping format - either online or at a traditional store. Figure 1 contains 

the hypothesized models to be tested in this study. 

-------------------------------- 

Figure 1 about here 

-------------------------------- 

As the dimensions of convenience have only begun to be examined empirically, the first 

hypothesis relates to the different structures of retail convenience for in-store and online shoppers. 

We concur with Seiders, Berry, and Gresham (2000) that there exists four distinct dimensions of 

retail convenience. For traditional in-store shoppers, we feel an additional dimension of 

convenience should be considered. Hoffman (2000) found that in-store shoppers consider it 

convenient to ask for and receive assistance from a salesperson in making a purchase, especially if 

the consumer lacks the necessary information to complete the purchase himself. Particularly for 

purchases that require extensive problem solving, if the consumer lacks the necessary information, 

he will seek that information from expert problem solvers like knowledgeable salespeople (Dunn, 

Thomas, and Lubawski 1981; Hoffman 2000). Thus, “assisted search,” defined as the speed and 

ease with which consumers identify and select products they wish to buy due to assistance from a 

salesperson, is an additional convenience dimension applicable to traditional retailers. It is not 

33

appropriate, however, to consider search features on a retailer’s website as assistance in the same 

manner in which it takes place at a physical store. It is the knowledge, training, and advice from a 

human salesperson that may provide convenience in a way that a virtual store cannot. Thus, the 

following is hypothesized: 

H1a: For traditional in-store shoppers, retail convenience is a multidimensional 

construct containing five distinct dimensions: access, search, assisted search, 

transaction, and possession. 

H1b: For online shoppers, retail convenience is a multidimensional construct 

containing four distinct dimensions: access, search, transaction, and possession. 

In addition to testing the multidimensional nature of retail convenience, we also hypothesize 

about the relationship of these dimensions to in-store and online shoppers. As previously stated, the 

dimension of access convenience is defined as “the speed and ease with which consumers can reach 

a retailer” (Seiders, Berry, and Gresham 2000, p. 81). Consumers who frequently shop at online 

retail stores do so because they can shop from the comfort of their homes and at any time of the day 

or night. The ability to reach the retailer at a time most convenient to the consumer (access 

convenience) is certainly a benefit of online shopping. Compared to shopping at traditional 

locations, shopping online saves the consumer travel time to the location, time spent parking, and 

time spent traveling from the parking lot to the store (Bhatnagar, Misra, and Rao 2000). With 

traditional locations, however, customers are required to adjust their preferred shopping time to fit 

within the retailer’s hours of operation. 

The concept of cost-benefit analysis (Prest and Turvey 1965) may be used to explain why 

consumers who shop online benefit most from access convenience. Online shoppers believe that 

the benefit of time saved by being able to access retail outlets from the comfort of their home or 

office at anytime of the day or night far outweighs the costs of delayed merchandise possession and 

the risks associated with shopping online (Wolfinbarger and Gilly 2001; Morganosky and Cude 

34

2000). In-store shoppers receive the benefit of immediate possession of their purchases at the 

sacrifice of convenient store access. Formally stated: 

H2: Online shoppers realize the benefit of access convenience more than traditional 

in-store shoppers do. 

Consumers who value search convenience do so because of reduced time spent looking for a 

particular product and finalizing their product decision. Online shoppers are more likely to realize 

the benefit of search convenience than in-store shoppers, simply because they are able to compare 

product and pricing information without having to leave home. Benefits falling within the 

dimension of search convenience include site design (Szymanski and Hise 2000), navigation 

(Childers, Carr, Peck, and Carson 2001), and selection and availability of product information 

(Wolfinbarger and Gilly 2001). Many consumers who choose to shop online do so because of the 

organization and design of the website and the ease of navigation. E-tailers having a user-friendly 

site design facilitate search convenience as consumers arriving at such a site can easily find exactly 

what they are looking for. Also considered in the dimension of search convenience is the product 

selection offered by the retailer. A virtual retailer is not limited by shelf space like a traditional 

retailer; therefore, they are usually able to offer a wider selection of products. The ease and speed 

with which the consumer can find exactly what he or she wants is key. The final benefit classified 

under search convenience is the availability of product information. How much information is 

available directly from the e-tailer’s virtual store? Do they supply comparisons to similar products 

offered by competing e-tailers? This is also important as many consumers turn to the Internet 

specifically for their information search. To conduct as extensive a search via traditional retailers 

would consume considerable amounts of time and effort. This leads to hypothesis 3: 

H3: Online shoppers realize the benefit of search convenience more than traditional 

in-store shoppers do. 

Customers seeking possession convenience prefer to save the time associated with actually 

obtaining the desired product (Seiders, Berry, and Gresham 2000). These consumers would rather 

35

spend time on other aspects of their shopping experience (e.g., getting to the store’s location and 

physically moving through the store to find exactly what they want) in exchange for not having to 

wait on parcel delivery to have their desired purchase in their possession. Applying cost-benefit 

analysis to this situation (Prest and Turvey 1965), consumers valuing possession convenience prefer 

to shop at traditional brick-and-mortar stores because the benefit of having the desired product in 

their hands at the end of the shopping trip outweighs the costs associated with traveling to the 

physical location and physically searching through the store’s shelves to find exactly what they 

want. In other words, they do not mind putting in extra shopping time as long as the desired 

product is in their possession immediately as a result of the shopping trip. Hypothesis 4 states the 

following: 

Construct Measurement 

H4: Traditional in-store shoppers realize the benefit of possession convenience 

more than online shoppers do. 

Research Method 

In order to measure the different dimensions of convenience, appropriate scale development 

procedures were followed (Churchill 1979; DeVellis 1991; Spector 1992). An initial survey 

containing several open-ended questions was administered to 196 students enrolled in upper-level 

marketing courses at a major Southeastern university. Questions such as “Please describe below 

what the word ‘convenience’ means to you” and “Describe as specifically as possible what your 

ideal convenient shopping experience would be like” were asked in order to develop the most 

appropriate phrases to capture each dimension of retail convenience. Many items were developed 

for each dimension, then 14 expert judges were asked to select the items that most closely matched 

the definitions of each dimension. A pretest was then conducted in order to improve/clarify 

question wording and instructions. These pretest surveys were given to a sample of seventy-five 

upper-level undergraduate marketing students to determine the necessary modifications. This 

36

process resulted in slightly different item wordings for in-store and online shoppers. Table 2 

contains the final set of items for each convenience dimension. 

Sampling Procedure 

-------------------------------- 


-------------------------------- 

The final survey version was administered to a convenience sample consisting of both 

students and non-students. Marketing students enrolled in upper-level undergraduate consumer 

behavior courses participated as both respondents and recruiters. Each student was asked to 

complete the survey and to recruit one other non-student to also complete the survey. Non-student 

names and phone numbers were collected and randomly checked to ensure authenticity. This 

process resulted in 308 total usable surveys (45% male, mean age 30); 241 completed their last 

major purchase in a traditional store, while 67 completed their last major purchase online. 

When answering the questions related to retail convenience, respondents were asked to think 

about their most recent major purchase. This was done to increase the chance that respondents 

would remember their purchase in as much detail as possible. Those respondents who made their 

most recent major purchase in a traditional store were directed to the scale items for in-store 

shopping, while those respondents who completed their most recent major purchase online were 

directed to the “online shopping” section of the survey. 

Statistical Technique 

Several statistical techniques were used to analyze the data. In order to test Hypothesis 1, 

statistics typically used in scale development were employed, including Cronbach’s alpha, principal 

components analysis, and confirmatory factor analysis using LISREL 8.3 for Windows. In order to 

test Hypotheses 2 through 4, analysis of variance was used to compare scale means of access, 

search, and possession convenience for in-store and online shoppers. 

37

Results 

In order to test Hypothesis 1, statistical procedures commonly used in scale development 

were employed. Reliabilities for each dimension of retail convenience were calculated using 

Cronbach’s alpha. Principal components analysis with varimax rotation was also undertaken in 

order to determine if five distinct dimensions were present for in-store shoppers, and four distinct 

dimensions were present for online shoppers. A more stringent confirmatory factor analysis using 

LISREL 8.3 for Windows was also undertaken in order to assess convergent and discriminant 

validity. Results for in-store shoppers are presented in Table 3, while results for online shoppers 

are presented in Table 4. Correlation matrices, means, and standard deviations of the scale items 

are provided for in-store shoppers in Table 5 and online shoppers in Table 6. 

-------------------------------- 


-------------------------------- 

-------------------------------- 


-------------------------------- 

-------------------------------- 


-------------------------------- 

-------------------------------- 


-------------------------------- 

Since all of the items within each dimension of retail convenience are reflective of their 

appropriate definitions, Cronbach’s alpha was used in order to examine the reliability of each 

dimension. As can be seen in the tables, these reliabilities were quite high, ranging from 0.7649 to 

0.9722. The dimensions of online retail convenience performed particularly well, as all were above 

the 0.9 level. 

For in-store shoppers, five dimensions of retail convenience were hypothesized to exist, 

while four dimensions of online retail convenience were hypothesized. In order to more carefully 

38

examine the dimensionality of these constructs, principal components analysis with varimax 

rotation was undertaken for each group separately. For in-store shoppers, five components with 

eigenvalues greater than one were extracted from the data. Together, these five components explain 

77.9% of the total variance. For online shoppers, four components with eigenvalues greater than 

one were extracted; together, they explain 84.4% of the total variance. Thus, there is strong 

evidence in support of Hypothesis 1; the dimensionality of retail convenience is different for the 

two different types of shopping formats. Evidence of discriminant validity is provided by the fact 

that these items did not want to load on other components with which they were not supposed to be 

associated. Further, evidence of convergent validity is provided by the high loadings of each item 

on their respective component. 

A more rigorous examination of the validity of these scales was conducted with 

confirmatory factor analysis using LISREL 8.3 for Windows. As can be seen in Tables 3 and 4, the 

statistically significant λ x parameter estimates provide evidence of convergent validity. 

Additionally, the majority of the squared multiple correlations (SMCs), defined as the percentage of 

variance in each item that is explained by the latent construct of interest, are above 50%, indicating 

that each item performed well in capturing its construct of interest. 

Modification indices associated with λ x and Θ δ reveal some problems with discriminant 

validity. Specifically, there were seven modification indices in the λ x matrix for in-store shoppers, 

indicating that items search2, help3, possess1, and possess2 wanted to be associated with other 

dimensions of convenience. Additionally, there were 13 modification indices in Θ δ that indicate 

some error terms wanted to correlate. These changes were not incorporated in the model, as 

theoretically it does not make sense to do so. These problematic modification indices explain the 

poor overall fit: chi-square=498.08, 198 df, p=0.00; RMSEA=0.08; RMR=0.06; GFI=0.85. 

Problematic modification indices also surfaced in the measurement model for online 

shoppers. Three items (access2, possess2, and possess4) wanted to be associated with other 

39

constructs, while 6 error terms wanted to correlate. Again, these changes were not implemented, 

resulting in poor overall model fit: chi-square=206.34, 97 df, p=0.00; RMSEA=0.13; RMR=0.09; 

GFI=0.71. The overall fit does substantially improve if paths are freed as suggested by the 

modification indices. Therefore, confirmatory factor analysis results are mixed; there is strong 

evidence of reliability and convergent validity, but discriminant validity must be improved in 

further scale refinement attempts. 

For Hypotheses 2 through 4, analysis of variance (ANOVA) was performed in order to 

compare the mean scale scores for online and in-store shoppers. Hypothesis 2 states that online 

shoppers benefit more from access convenience than in-store shoppers. A comparison of scale 

means for these two groups yields an F statistic of 37.59 (p=0.000). Access convenience is more 

beneficial for online shoppers (scale mean=6.55) than in-store shoppers (scale mean=5.65), 

providing evidence of support for H2. Hypothesis 3 states that online shoppers benefit more from 

search convenience than do in-store shoppers; ANOVA here produces an F statistic of 8.57 

(p=0.004). The scale mean of 5.88 for in-store shoppers is significantly less than the scale mean of 

6.26 for online shoppers; thus, there is evidence of support for H3. Finally, hypothesis 4 predicts 

that in-store shoppers benefit from possession convenience more than online shoppers. The F 

statistic from this ANOVA is significant (F=11.07, p=0.001); however, the scale means indicate 

that the reverse is true. The possession scale mean for in-store shoppers is 5.60, while the scale 

mean for online shoppers is 6.23. This finding is quite counter-intuitive; possible reasons for this 

result are provided in the next section. 

Discussion 

This study provides an important contribution to the marketing convenience literature 

because it is the first of its kind to utilize the different dimensions of retail convenience to better 

understand customer needs. The support for Hypothesis 1 suggests that the dimensions of 

convenience are different for the two different types of shopping formats. The expansion of 

40

Seiders, Berry, and Gresham’s (2000) four dimensions of retail convenience to include “assisted 

search” is warranted. The five-component solution resulting from the principal components 

analysis is evidence of the existence of this newly proposed dimension. Further work in this area 

should seek to establish a similar convenience dimension for online shoppers. Technological 

capabilities that allow customers to experience “live help” or to receive recommendations based on 

what he has in his shopping bag may be considered a different type of “assisted help” than what is 

provided in a traditional store. It may be difficult, however, to obtain a large enough sample of 

online shoppers who regularly use these convenience-oriented features. 

The scale items crafted to measure the different dimensions of retail convenience possess 

high reliability levels and strong evidence of convergent validity as shown by Cronbach’s alpha and 

statistically significant λ x parameter estimates respectively. A more stringent confirmatory factor 

analysis using structural equation modeling reveals problems with discriminant validity, as 

evidenced by large λ x and Θ δ modification indices. Some scale items want to be associated with 

constructs that they are not supposed to be associated with. In addition, some error terms want to 

correlate with other error terms. This means that insufficient evidence exists to suggest that the 

dimensions of retail convenience are in fact different dimensions. Further studies must be 

undertaken to ensure that the dimensions of access, search, assisted search, transaction, and 

possession convenience are all distinct dimensions of convenience. 

The support of Hypothesis 2 suggests that online shoppers do benefit from access 

convenience more than traditional in-store shoppers do. One implication of this finding is that 

when given a choice between the two retail formats (online shopping or traditional in-store 

shopping) customers who value access convenience above all other types of convenience will often 

choose to shop online. Hypothesis 3 (that online shoppers benefit from search convenience more 

than traditional in-store shoppers) is also supported. When given a choice between the two retail 

41

formats, customers who value search convenience above all other types of convenience will also 

choose to shop online. 

From a marketing standpoint, the results of Hypothesis 2 and 3 suggest that retailers should 

try to better understand their customers; specifically, they should try to find out what type of 

convenience their customers want. It is possible that customers may value multiple convenience 

dimensions. For example, a customer may value both search convenience and possession 

convenience. One way to better satisfy this customer is to offer multiple channel outlets (maintain 

both a traditional brick-and-mortar store and an online store). By having multiple channel outlets, 

the retailer facilitates both search convenience and possession convenience. In other words, the 

customer can use the online store for his information search, then go to the traditional store and 

leave with his desired purchase. Offering multiple channels to reach customers offers ultimate 

convenience. 

Upon further review of hypothesis 4, the seemingly counter-intuitive results can be easily 

explained. It is true that online shoppers must wait longer for their purchases than in-store 

shoppers. But a post hoc examination of the item wordings that measured possession convenience 

for online shoppers reveals that this wait was not adequately captured. Instead, we asked if the 

order was delivered in a timely fashion. This captures satisfaction with possession, rather than how 

conveniently the possession actually took place. In subsequent scale refinement attempts, 

possession convenience might be captured more precisely if an objective measure was used. 

Suppose the question was worded, “How long did it take for you to receive the product(s) you 

ordered?” An in-store shopper would answer “zero” (the most convenient possession), and the 

longer one had to wait for his online order, the less convenient his possession would be. 

Conclusion 

This study is an initial attempt to delineate the different retail convenience dimensions for 

in-store and online shoppers. Several limitations must be mentioned. A convenience sample was 

42

used, and the respondents were all concentrated in one geographical area. Additionally, we 

obtained a small sample size of online shoppers. Scale refinement will greatly benefit from a 

national sample with equally large numbers of in-store and online shoppers. 

As mentioned above, future research should aim to establish discriminant validity of the 

convenience scales. Additional comparisons of in-store and online shoppers are also warranted. 

Future studies should investigate the time and effort expenditures for these two groups as well as 

examine possible antecedents to convenience-seeking behaviors. Constructs such as time pressure 

and shopping orientation may lead consumers to seek different types of convenience. Retailers 

would greatly benefit from such an understanding of their customers’ needs and motivations. 

43

References 

Anderson Jr. WT. Convenience orientation and consumption behavior. Journal of Retailing 

1972;48(3):49-72. 

Berry LL. The time-buying consumer. Journal of Retailing 1979;55(4):58-69. 

Berry LL, Seiders K, Grewal D. Understanding service convenience. Journal of Marketing 

2002;66(3):1-17. 

Berry LL. The old pillars of new retailing. Harvard Business Review 2001;79(4):131-137. 

Bettman JR, Johnson EJ, Payne JW. A componential analysis of cognitive effort in choice. 

Organizational Behavior and Human Decision Processes 1990;45(February):111-139. 

Bhatnagar A, Misra S, Rao HR. On risk, convenience, and Internet shopping behavior. 

Communications of the ACM 2000;43(11):98-110. 

Brown LG. The strategic and tactical implications of convenience in consumer product marketing. 

Journal of Consumer Marketing 1989;6(3):13-19. 

Brown LG. Convenience in services marketing. Journal of Services Marketing 1990;4(1):53-59. 

Bucklin LP. Retail strategy and the classification of consumer goods. Journal of Marketing 

1963;27(1):50-55. 

Burke RR. Technology and the customer interface: what consumers want in the physical and 

virtual store. Journal of the Academy of Marketing Science 2002;30(4):411-432. 

Childers T L, Carr CL, Peck J, Carson S. Hedonic and utilitarian motivations for online retail 

shopping behavior. Journal of Retailing 2001;77(4):511-535. 

Churchill GA. A paradigm for developing better measures of marketing constructs. Journal of 

Marketing Research 1979;16(1):64-73. 

Copeland MT. Relation of consumers’ buying habits to marketing methods. Harvard Business 

Review 1923;1(3):282-289. 

Crist B. Myths about convenience food costs. Journal of Marketing 1960;24(4):49-55. 

Darian JC, Cohen J. Segmenting by consumer time shortage. Journal of Consumer Marketing 

1995;12(1):32-44. 

DeVellis RF. Scale development: theory and applications (Applied Social Research Methods 

Series, Vol. 26). Newbury Park: Sage Publications, 1991. 

Dunn DT, Thomas CA, Lubawski JL. Pitfalls of consultative selling. Business Horizons 

1981;24(5):59-65. 

44

Enis BM, Roering KJ. Product classification taxonomies: synthesis and consumer implications. In: 

Theoretical developments in marketing. Chicago: American Marketing Association, 1980. 

p. 186-189. 

Etgar, M. The household as a production unit. In: Sheth JN, editor. Research in marketing, vol. 1. 

Greenwich, CT: JAI Press, 1978. p. 79-98. 

Foote NN. The Image of the Consumer in the Year 2000. Proceedings, Thirty-fifth Annual Boston 

Conference on Distribution 1963, 13-18. 

Forster J. Hungry for convenience. Business Week (1/14/2002); Issue 3765: 120. 

Gardner EH. Consumer goods classification. Journal of Marketing 1945;9(3):275-276. 

Green H. Happy holidays – for e-tailers, at least. Business Week (11/11/02); Issue 3807:44-45. 

Gross BL. Time scarcity: interdisciplinary perspectives and implications for consumer behavior. 

In: Sheth JN, Hirschman EC, editors. Research in consumer behavior. Greenwich, CT: JAI 

Press, 1987. p. 1-54. 

Hof RD. Don’t cut back now. Business Week (10/1/2001); Issue 3751:EB34. 

Hofacker CF. Internet Marketing, 3 rd ed. New York: John Wiley & Sons, Inc., 2001. 

Hoffman NP. May I help you?: customer intimacy versus customer empowerment in a retail context. 

Presented at the Third Annual Retail Strategy and Consumer Decision Research Seminar 2000, 

Orlando, FL, November. 

Holbrook MB, Howard JA. Frequently purchased nondurable goods and services. In: Ferber R, 

editor. Selected aspects of consumer behavior: a summary from the perspective of different 

disciplines. Washington D.C.: National Science Foundation, Directorate for Research 

Applications, Research Applied to National Needs, 1977. p. 189-222. 

Jacoby J, Szybillo GJ, Berning CK. Time and consumer behavior: an interdisciplinary overview. 

Journal of Consumer Research 1976;2(4):320-339. 

Kaish S. Cognitive dissonance and the classification of consumer goods. Journal of Marketing 

1967;31(4):28-31. 

Kaufman-Scarborough C. , Lindquist JD. E-shopping in a multiple channel environment. Journal 

of Consumer Marketing 2002; 19(4):333-350. 

Kelley EJ. The importance of convenience in consumer purchasing. Journal of Marketing 

1958;23(1):32-38. 

45

Kotler P, Zaltman G. Social marketing: an approach to planned social change. Journal of 

Marketing 1971;35(3):3-12. 

Lovelock CH. Classifying services to gain strategic marketing insights. Journal of Marketing 

1983;47(3):9-20. 

Morganosky MA, Cude BF. Consumer response to online grocery shopping. International Journal 

of Retail and Distribution Management 2000;28(1):17-26. 

Murphy PE, Enis BM. Classifying products strategically. Journal of Marketing 1986;50(3):24-42. 

Prest AR, Turvey R. Cost-benefit analysis: a survey. Economic Journal 1965;75(300):683-735. 

Reilly MD. Working wives and convenience consumption. Journal of Consumer Research 

1982;8(4):407-418. 

Seiders K, Berry LL, Gresham LG. Attention, retailers! How convenient is your convenience 

strategy? Sloan Management Review 2000;41(3):79-89. 

Seiders K, Voss GB, Grewal D, Godfrey AL. Customer evaluation of service convenience: an 

empirical investigation. Proceedings, American Marketing Association Winter Educators’ 

Conference, (Winter 2003), 163-164. 

Shop.org. The multi-channel retail report 2001. Washington D.C.: National Retail Federation, 

2001. 

Spector, PE. Summated Rating Scale Construction: An Introduction. Newbury Park: Sage 

Publications, 1992. 

Szymanski DM, Hise RT. E-satisfaction: an initial examination. Journal of Retailing 

2000;76(3):309-322. 

Wolfinbarger M, Gilly MC. Shopping online for freedom, control, and fun. California 

Management Review 2001;43(2):34-55. 

Youngdahl WE, Kellog DL. The relationship between service customers’ quality assurance 

behaviors, satisfaction, and effort: a cost of quality perspective. Journal of Operations 

Management 1997;15(1):19-32. 

46

Table 1 

Dimensions of Convenience 

General Convenience Definition 

(Brown 1989, 1990) 

1. Time Products/services may be provided at a time that is most 

convenient for the customer 

2. Place Products/services may be provided in a place that is more 


3. Acquisition Firms may make it easier for the customer, financially and 

otherwise, to purchase their products/services 

4. Use Products/services may be made more convenient for the 

customer to use 

5. Execution Perhaps the most obvious convenience is simply having some 

one else provide the product/service for the consumer 

Service Convenience (Berry, 

Seiders, and Grewal 2002) 

1. Access convenience Involves consumers’ perceived time and effort expenditures to 

initiate service delivery 

2. Decision convenience Involves consumers’ perceived time and effort expenditures to 

make service purchase or use decisions 

3. Transaction convenience Involves consumers’ perceived expenditures of time and effort 

to effect a transaction 

4. Benefit convenience Involves consumers’ perceived time and effort expenditures to 

experience the service’s core benefits 

5. Post-benefit convenience Involves consumers’ perceived time and effort expenditures 

when reinitiating contact with a firm after the benefit stage of 

the service 

Retail Convenience (Seiders, 

Berry, and Gresham 2000) 

1. Access convenience The speed and ease with which consumers can reach a retailer 

2. Search convenience The speed and ease with which consumers can identify and 

select products they wish to buy 

3. Transaction convenience The speed and ease with which consumers can effect or amend 

transactions 

4. Possession convenience The speed and ease with which consumers can obtain desired 

products 

47

Dimension Items for In-Store Shoppers 1 

Table 2 

Scale Items Measuring Retail Convenience 

Items for Online Shoppers 1 

Access 1. The store was easy to get to. 1. The website was easy to find. 

2. The store had convenient hours. 2. I could order any time I wanted. 

3. Parking was reasonably available. 3. I could order from remote locations 

4. I was able to get to the store’s location 

quickly. 

(e.g., home, work, etc.) 

4. I was able to find the website quickly. 

Search 1. The store was well-organized. 1. The website was well-organized. 

2. I could easily find what I was looking 2. It was easy to find what I was looking 

for. 

for. 

3. The store was neat. 3. It was easy to navigate the website. 

4. The store was clean. 4. The website provided useful 

information. 

5. It was easy to get the information I 

needed to make my purchase decision. 

Help with Search 1. A salesperson was readily available to 

help me. 

2. The salesperson offered good advice. 

3. The salesperson helped me find exactly 

what I was looking for. 

4. The salesperson was knowledgeable. 

5. The salesperson gave me useful 

information. 

6. I thought the salesperson was welltrained. 

Transaction 1. The store has a fast checkout. 1. The checkout process was fast. 

2. My purchase was completed easily. 2. My purchase was completed easily. 

3. I was able to complete my purchase 3. It didn’t take a long time to complete 

quickly. 

4. I didn’t have to wait to pay. 

5. It didn’t take a long time to complete 

the purchase process. 


Possession 1. I got exactly what I wanted. 1. I got exactly what I wanted. 

2. It took a minimal amount of effort on 2. It took a minimal amount of effort on 

my part to get what I wanted. 


3. I got what I wanted when I wanted it. 3. My order was delivered in a timely 

fashion. 

4. I was properly notified of my order 

status. 

1All items were measured on a 7-point scale anchored by Strongly Agree and Strongly Disagree. 

48

Table 3 

Retail Convenience Scale Results for In-Store Shoppers 

Dimensio 

n 

Alpha Component CFA Results 

1 2 3 4 5 λx estimate t-value SMC 

Access1 0.7759 0.87 

7 

0.89 16.56 0.80 

Access2 .077 

1 

0.73 12.48 0.53 

Access3 .063 

1 

0.57 9.26 0.33 

Access4 0.85 

7 

0.79 13.95 0.62 

Search1 0.8860 0.72 

6 

Search2 0.62 

2 

Search3 0.85 

8 

Search4 0.82 

6 

Help1 0.9503 0.77 

3 

Help2 0.88 

1 

Help3 0.79 

9 

Help4 0.92 

3 

Help5 0.90 

5 

Help6 0.87 

1 

Transact1 0.9609 0.88 

5 

Transact2 0.88 

7 


1 


0 


3 

0.84 15.27 0.71 

0.76 13.28 0.58 

0.78 13.62 0.61 

0.73 12.43 0.53 

0.75 13.74 0.56 

0.90 18.15 0.81 

0.84 16.26 0.71 

0.94 19.71 0.89 

0.94 19.45 0.88 

0.88 17.33 0.77 

0.87 17.24 0.76 

0.95 20.05 0.90 

0.98 21.07 0.95 

0.84 16.27 0.70 

0.95 20.19 .091 

Possess1 0.7649 0.761 0.76 12.39 0.57 

Possess2 0.663 0.67 10.74 0.45 

Possess3 0.778 0.72 11.63 0.52 

n=241 

49

Dimensio 

n 

Table 4 

Retail Convenience Scale Results for Online Shoppers 


1 2 3 4 λx estimate t-value SMC 

Access1 0.9298 0.866 0.86 8.26 0.73 

Access2 0.735 0.81 7.65 0.66 

Access3 0.852 0.93 9.59 0.87 

Access4 0.903 0.92 9.34 0.85 

Search1 0.9033 0.537 0.86 8.31 0.75 

Search2 0.508 0.87 8.46 0.76 

Search3 0.738 0.88 8.65 0.78 

Search4 0.866 0.66 5.61 0.43 

Search5 0.850 0.65 5.55 0.42 

Transact1 0.9722 0.921 0.96 10.22 0.93 

Transact2 0.914 0.97 10.35 0.94 

Transact3 0.918 0.95 10.04 0.91 

Possess1 0.9110 0.800 0.85 8.12 0.73 

Possess2 0.669 0.83 7.80 0.69 

Possess3 0.845 0.87 8.45 0.76 

Possess4 0.897 0.84 7.98 0.71 

n=67 

50

Table 5 

Correlations, Means, and Standard Deviations of Scale Items for In-store Shoppers 

Item A1 A2 A3 A4 S1 S2 S3 S4 H1 H2 H3 H4 H5 H6 T1 T2 T3 T4 T5 P1 P2 P3 

Acc1 1 

Acc2 .63 1 

Acc3 .49 .50 1 

Acc4 .73 .54 .40 1 

Sear1 .37 .41 .40 .31 1 

Sear2 .26 .29 .25 .21 .63 1 

Sear3 .23 .32 .30 .16 .67 .60 1 

Sear4 .27 .35 .34 .21 .63 .53 .80 1 

Help1 .20 .12 .10 .18 .38 .38 .38 .33 1 

Help2 .17 .14 .06 .10 .39 .41 .39 .35 .70 1 

Help3 .18 .17 .12 .11 .41 .48 .42 .37 .68 .81 1 

Help4 .19 .17 .07 .10 .40 .38 .38 .38 .70 .83 .78 1 

Help5 .18 .15 .07 .08 .35 .36 .39 .37 .66 .85 .75 .90 1 

Help6 .15 .11 .06 .14 .36 .40 .41 .38 .69 .76 .72 .84 .82 1 

Tran1 .21 .28 .23 .12 .37 .33 .22 .31 .25 .31 .35 .27 .26 .22 1 

Tran2 .21 .24 .22 .13 .33 .43 .33 .34 .29 .34 .42 .31 .29 .32 .83 1 

Tran3 .18 .22 .23 .11 .32 .42 .32 .32 .26 .31 .39 .27 .27 .27 .84 .93 1 

Tran4 .12 .12 .22 .03 .31 .38 .26 .36 .27 .28 .35 .28 .23 .25 .79 .77 .81 1 

Tran5 .18 .20 .17 .10 .29 .41 .30 .29 .26 .32 .39 .28 .25 .25 .83 .90 .93 .82 1 

Pos1 .07 .10 .03 .05 .27 .39 .27 .22 .39 .46 .56 .42 .44 .43 .30 .41 .36 .30 .37 1 

Pos2 .27 .24 .20 .16 .35 .36 .31 .30 .23 .30 .34 .20 .25 .21 .41 .46 .49 .40 .46 .49 1 

Pos3 .21 .25 .16 .13 .35 .42 .32 .33 .33 .29 .39 .24 .30 .30 .32 .39 .37 .29 .32 .57 .47 1 

Mean 5.6 5.9 5.8 5.3 5.8 5.8 5.9 6.0 6.0 5.7 5.8 6.0 5.8 5.9 5.1 5.3 5.1 5.2 5.1 6.0 5.0 5.8 

SD 1.4 1.2 1.3 1.7 1.1 1.2 1.2 1.0 1.4 1.5 1.5 1.3 1.3 1.3 1.5 1.6 1.6 1.7 1.7 1.3 1.8 1.5 

n=241 

Table 6 

Correlations, Means, and Standard Deviations of Scale Items for Online Shoppers 

Item A1 A2 A3 A4 S1 S2 S3 S4 S5 T1 T2 T3 P1 P2 P3 P4 

Acc1 1 

Acc2 .64 1 

Acc3 .77 .79 1 

Acc4 .85 .72 .86 1 

Sear1 .56 .66 .62 .59 1 

Sear2 .51 .67 .67 .55 .76 1 

Sear3 .47 .66 .60 .52 .78 .74 1 

Sear4 .29 .51 .38 .37 .52 .51 .71 1 

Sear5 .29 .46 .33 .28 .48 .59 .67 .77 1 

Tran 

1 

.24 .53 .34 .33 .43 .58 .48 .38 .42 1 

Tran 

2 

.16 .48 .26 .21 .39 .55 .49 .46 .54 .93 1 

Tran .15 .42 .27 .22 .37 .50 .48 .43 .50 .92 .93 1 

3 

Pos1 .44 .45 .46 .42 .62 .65 .50 .30 .42 .54 .47 .43 1 

Pos2 .41 .63 .56 .48 .60 .66 .51 .41 .43 .61 .53 .55 .74 1 

Pos3 .47 .47 .59 .52 .51 .59 .49 .20 .19 .38 .29 .29 .71 .70 1 

Pos4 .41 .35 .51 .43 .53 .48 .41 .19 .16 .31 .24 .24 .72 .65 .81 1 

Mean 6.5 6.6 6.5 6.6 6.2 6.3 6.2 6.3 6.3 5.7 5.7 5.6 6.4 5.9 6.2 6.4 

SD .89 .70 .80 .82 .91 .98 .91 .85 .88 1.2 1.4 1.3 .80 1.2 1.2 1.1 

51

n=67 

52

Error! 

Figure 1 

Retail Convenience for In-Store and Online Shoppers 

In-Store Retail 

Convenience 

Access Search Assisted 

Search 

Transaction Possession 

Online Retail 

Convenience 

Access Search Transaction Possession 

53

A study of the impact of shopping orientation and gender on the value-satisfaction link during 

an electronic catalog visit 

Christine Gonzalez, Toulouse Business School 

Abstract 

This paper studies how shopping orientation – shopping as work vs. shopping as recreation- and 

gender modify the links between online shopping values and consumer satisfaction. An experience was 

conducted in which shopping orientation was manipulated. 417 students visited an electronic catalog with 

“shopping as recreation” scenario and another electronic catalog with a “shopping as work” scenario. The 

results lend support to the hypothesized moderating effects. Implications for e-marketers as well as 

directions for future research are discussed. 

Introduction 

Consumer satisfaction with online retailing has been recently investigated (Szymanski and Hise, 

2000; Zeithaml, Parasuraman and Malhotra, 2000; Zeithaml, Parasuraman and Malhotra, 2002). However, 

although consumer motives shape consumer responses to the traditional or online shopping environment 

(Bloch, Sherrell and Ridgway, 1986; Bloch, Ridgway and Sherrell, 1989; Hoffman and Novak, 1996; 

Hammond, McWilliam and Diaz, 1998), few studies consider their role in the online satisfaction process. A 

traditional or online store experience is evaluated on its ability to produce both hedonic and utilitarian 

outcomes (Babin, 1991; Babin, Darden and Griffin, 1994; Griffin, Babin and Modianos, 2000; Babin and 

Attaway, 2001; Mathwick, Malhotra and Ridgon, 2001; Senecal, Gharbi and Nantel, 2001). As the Internet 

has been depicted as a convenient shopping option (Szymanski and Hise, 2000; Seiders, Berry and Gresham, 

2000), it rises the question of the importance of the hedonic outcomes of the online experience in 

determining consumer satisfaction. Should online retailers focus only on efficiency or on both hedonic and 

utilitarian aspects of the shopping experience? How do online shopping values influence consumer 

satisfaction? 

Although consumer situation shape consumer responses to the traditional or online shopping 

environment (Bloch, Sherrell and Ridgway, 1986; Bloch, Ridgway and Sherrell, 1989; Hoffman 

55

and Novak, 1996; Hammond, McWilliam and Diaz, 1998), few studies consider its role in the 

online satisfaction process. Product search, information search and recreation are common motives 

for surfing on the Internet (Hoffman and Novak, 1996; Hammond, McWilliam and Diaz, 1998). 

Purchasing conversion rate is only five per cent on the Internet (Moe and Fader, 2001). However all 

visitors-information seekers or recreation seekers- are valuable to an online retailer: as a shopper 

makes more visits to a store, he/she feels less anxiety toward buying on the web site (Moe and 

Fader, 2001). In addition to shopping motives others dimensions of consumer situation such as time 

pressure also influence consumer shopping behaviour and its response to the shopping environment 

(Iyer, 1989; Titus and Everett, 1997). Understanding consumer situation while e-shopping is a key 

requirement to build online retail patronage: How does consumer situation shape consumer 

shopping experience and the satisfaction process itself? Based on a synthesis of previous research, 

two specific shopping orientations have been defined here: a “shopping as work” orientation (“the 

consumer considers the visit ahead more as a job that has to be done and is oriented toward an 

efficient experience that accomplishes the job”) and a “shopping as recreation” orientation (“the 

consumer considers the visit ahead as being potentially pleasant and expects to gain gratification 

from the process itself"). What role do online shopping orientations play in shaping the satisfaction 

process? 

Few researchers try to understand how males and females differ in their evaluation of a traditional or 

online shopping experience. However, because of socio-demographical changes- such as an increase in 

divorce rate in most western countries- or behavioural changes- men in their teens or mid-20s are, for 

example, shop more frequently and are beginning to shop in group like teenage girls (Bird, 2002), males are 

becoming an important part of the store clientele. At the end of 2002, in the United States, males represented 

51% of all internet users (Harris Interactive cited by Journal du Net, 2003) and 40% of Internet shoppers 

(BizRate cited by Journal du Net, 2003). It is thus important to understand how they react to an online 

56

shopping environment. How do males and females differ in their evaluation of an online shopping 

experience? 

In order to answer these questions, this article starts with a conceptual discussion of online 

shopping values, satisfaction, shopping orientation and gender. Literature in consumer behavior 

with a focus on shopping behavior is used to understand the relationships between these variables. 

The results of an online experiment which validates some of these relationships are then presented. 

Finally, implications for theory and practice are addressed. 

Online shopping values and consumer satisfaction with the electronic catalog visit 

Shoppers’ experiences in connection to an online or traditional store can be categorized into 

(1) experiences related to being in the store itself and (2) experiences related to consuming the 

products and services obtained from the retailer (Westbrook, 1981). Online store satisfaction results 

from shopping experience satisfaction, product delivery satisfaction, product satisfaction and when 

necessary complaining behavior satisfaction (Oliver, 1981). This article focuses on the first step of 

the process-the electronic catalog visit- and on consumer satisfaction with the electronic catalog 

visit. Satisfaction has been defined here as “an evaluation of the surprise inherent in a product 

acquisition and/or product experience” (Oliver, 1981). 

Consumption experiences have been described according to an information processing and 

an experiential perceptive (Holbrook and Hirschman, 1982; Hirschman and Holbrook, 1982). The 

information processing perspective describes the consumer as “a problem solver engaged in the 

goal directed activities of searching for information, retrieving memory cues, weighting evidence, 

and arriving at carefully considered judgmental evaluation” (Holbrook and Hirschman, 1982, p 

135) while the experiential view describes the consumer as an hedonist engaged in “playful leisure 

activities, sensory pleasures, daydreams, esthetic enjoyment, and emotional responses” (Holbrook 

57

and Hirschman, 1982, p 132). These two perspectives are complementary rather than contradictory 

(Holbrook, 1986; Babin, 1991; Holbrook, 1994; Babin, Darden and Griffin, 1994 Holbrook, 1999). 

As a matter of fact, a service or store experience is evaluated on both its hedonic and its utilitarian 

outcomes (Babin, 1991; Mano and Elliott, 1993; Babin, Darden and Griffin, 1994; de Ruyter, 

Wetzels, Lemmink and Mattsson, 1997; Lemmink, de Ruyter and Wetzels, 1998; Raman and 

Leckenby, 1998; Griffin, Babin and Modianos, 2000). In the e-retailing context, value is derived 

from Web sites that combine an efficient visit and excellent service quality with a fun and playful 

experience and an aesthetic shopping environment (Mathwick, Malhotra and Rigdon, 2001). 

Past research on patronage behavior made a distinction between shopping hedonic and 

utilitarian value (Babin, 1991; Mano and Elliott, 1993; Babin, Darden and Griffin, 1994; Griffin, 

Babin and Modianos, 2000). Hedonic value results from the recreation and pleasure provided by the 

shopping experience while utilitarian value is realized when the task is completed efficiently 

(Babin, Darden and Griffin, 1994; Griffin, Babin and Modianos, 2000). Hedonic and utilitarian 

values are antecedents of consumer satisfaction with the service or shopping experience (Babin, 

Darden and Griffin, 1994; de Ruyter, Wetzels, Lemmink and Mattsson, 1997; Lemmink, de Ruyter 

and Wetzels, 1998; Babin and Attaway, 2001). 

« Shopping as work » versus « Shopping as Recreation » Orientation 

A variety of hedonic and utilitarian motives has been suggested in the retailing literature (Tauber, 

1972; Bellenger, Robertson and Greenberg, 1977; Bellenger and Korgaonkar, 1980; Westbrook and Black, 

1985; Arnold and Reynolds, 2003). Tauber (1972) identified personal (i.e. role playing, diversion, self- 

gratification, learning about new trends, physical activities and sensory stimulation) and social (i.e. social 

experiences outside the home, communication with others having a similar interest, peer group attraction, 

status and authority, pleasure of bargaining) motives for shopping. Later, Westbrook and Black (1985) added 

two more utilitarian motives to these hedonic motives for shopping: anticipated utility and choice 

58

optimization. More recently, Arnold and Reynolds (2003) provided a comprehensive inventory of 

consumers’ hedonic shopping motivations and have isolated adventure, gratification, role, value, social and 

idea shopping motivations. Westbrook and Black (1985) summarize these taxonomies of shoppers into three 

categories: product oriented motives- to acquire the product for which needs are experienced or to acquire 

product information, a combination of product oriented and experiential motives- both to acquire the desired 

product and to have a pleasurable recreational experience- and experiential motives- principally to have a 

pleasurable recreational experience. 

Shopping motives shape the shopping experience (Dawson, Bloch and Ridgway, 1990). For 

example, product oriented motives are associated with greater pleasure than arousal while experiential 

motives are linked with greater arousal than pleasure. Product oriented motives are a significant predictor of 

most dimensions of consumer preference including expectations met, product satisfaction and overall 

satisfaction. Experiential motives significantly predict facility satisfaction and overall satisfaction. Shopping 

motives should also shape consumer behavior online. For example, task completion and prepurchase 

deliberation motives should induce utilitarian and instrumental goal directed behavior and recreational 

motives should encourage a more ritualistic, transcendental, hedonic behavior (Hoffman and Novak, 1996). 

However Hammond, McWilliam and Diaz (1999) - who defined web browsing for general information as 

hedonic and web browsing for specific information as utilitarian - did not find relationships between motives 

and hedonic and utilitarian attitudes toward the web. 

Motives are however only one dimension of consumer situation while visiting the online or 

traditional store. As a matter of fact, Belk (1975) postulated that an interaction between an object 

and a situation determines consumer internal and behavioral responses. He isolated four dimensions 

of consumer situation in addition to motives: physical surroundings, social surroundings, temporal 

perspective and antecedent states. According to Lutz and Kakkar (1975; p. 440), “how the actor 

perceives the situation is as important as the actual elements found in the physical environment.” 

Thus the situation relevant for the understanding of consumer behavior is the psychological 

situation defined by Lutz and Kakkar (1975; p. 441) as “an individual’s internal responses to, or 

59

interpretations of, all factors particular to a time and place of observation.” Internal responses are 

captured by the three emotion dimensions: Pleasure, Arousal and Dominance (PAD) (Mehrabian 

and Russell 1974). Additionally, the consumer has expectations about things including efficiency 

and recreation that they use to judge a traditional or online shopping experience (Titus and Everett, 

1995). As a consequence, the subjective situation is framed by consumer expectations: hedonic 

expectations such as recreation, fun or pleasure and utilitarian expectations such as efficiency and 

accomplishment. 

Based on these researches we contrast a “shopping as work” and a “shopping as recreation” 

orientation. In the “shopping as work” orientation (SWO), before visiting the web site, the 

consumer does not consider the visit as being potentially gratifying in and of itself and is oriented 

more toward efficiency and task completion. In the “shopping as recreation” orientation (SRO), 

before visiting the web site, the consumer consider the visit as being potentially gratifying in and of 

itself and is oriented more toward recreation. 

Objective situation 

Subjective situation 

Shopping orientation 

Physical surroundings 

Social surroundings 

Temporal perspective 

Antecedent states 

Motives 

Emotional states 

Expectations 

Figure 1 – Shopping as work vs. Shopping as recreation orientation 

Shopping orientation shapes the satisfaction process and modifies relationships between the 

store environment and the evaluation of the shopping experience. For example, consumer 

satisfaction with a restaurant depends mainly on utilitarian value for a business lunch and mainly on 

hedonic value for a celebration meal (Lemmink, de Ruyter and Wetzels, 1998). Thus, 

60

H1 : 

H2 : 

The relationship between online shopping hedonic value and satisfaction will be 

stronger for a “shopping as recreation” orientation 

The relationship between online shopping utilitarian value and satisfaction will be 

stronger for a “shopping as work” orientation 

Hammond, McWilliam and Diaz (1998) posit that consumer motives (browsing vs finding a precise 

information) have an impact on the intensity of one's hedonic and utilitarian attitude toward the 

web. 

H3 : 

H4 : 

Online shopping hedonic value should be stronger for a “shopping as recreation” 

orientation 

Online shopping utilitarian value should be stronger for a “shopping as work” 

orientation 

Gender differences and the online shopping experience 

How do males and females evaluate a shopping experience? According to Ezell and Motes 

(1995), in a traditional store, males are more sensible than females to store characteristics that 

enable them to carry out the grocery shopping activity in minimum time: ease of finding desired 

items, convenience of location and ease of navigation inside the store. As a matter of fact, even 

when they engage in shopping willingly and enjoy this activity, males are goal-directed and “shop 

to win” (Otnes and McGrath, 2001). They shop to achieve shopping success, for example to defeat 

retailers by obtaining lower prices, status, for example by demonstrating expertise in certain product 

categories, or romantic success, by mastering feminine product categories such as lingerie or 

gourmet food for example (Otnes and McGrath, 2001). Gender differences in the evaluation of 

shopping should also be true in the context of e-shopping. Males are indeed more sensible than 

females to the instrumental aspects of new technologies: they are task-oriented and consider 

perceived usefulness to a greater extent in making their decisions regarding the adoption of a new 

technology (Venkatesh and Morris, 2000). As a consequence, females should be more sensible to 

61

the hedonic aspects of the online shopping experience and males should be more sensible to its 

utilitarian aspects. 

H 5 : The relationship between online shopping hedonic value and satisfaction will be 

stronger for females than males 

H 6 : The relationship between online shopping utilitarian value and satisfaction will be 

stronger for males than for females 

H 7 : Online shopping hedonic value should be stronger for females 

H 8 : Online shopping utilitarian value should be stronger for males 

Online shopping 

hedonic value 

Online shopping 

utilitarian value 

Research methods 

Experience 

Shopping orientation 

Gender 

Figure 2 – Conceptual framework 

Satisfaction 

An experiment manipulated the shopping purpose over two levels. 417 undergraduate and 

graduate business students participated in the study. Four electronic catalogs from online travel 

agencies were selected as stimuli for the study: online travel agencies were selected because travel 

is among the most successful product on the web. In order to test the links between online shopping 

values and consumer satisfaction, each subject had to visit two electronic catalogs and then respond 

to a questionnaire. To avoid any bias, subjects who had already visited the electronic catalog were 

excluded from the sample, the final sample is of 322 students, 142 females and 180 males. 

62

Each subject experience both experimental conditions and visit an electronic catalog with a 

“shopping as work” scenario and another catalog with a “shopping as recreation” scenario. It 

guarantees that differences between the two scenarios are caused by the manipulation, and not by 

sampling differences. Confounding variables such as carryover effects were controlled by 

counterbalancing. Half the subjects visited first an electronic catalog with a “shopping as 

recreation” orientation and then another one with a “shopping as work” orientation and the other 

half visited first an electronic catalog with a “shopping as work” orientation and then another one 

with a “shopping as recreation” orientation. 

A scenario was developed for each orientation and for each electronic catalog. A pre-test was 

performed to check that the “shopping as work” scenario was perceived as work and that the 

“shopping as recreation” scenario was perceived as recreation. Additionally a manipulation check 

was introduced in the final questionnaire: “How did you consider the visit before visiting the web 

site: as work � as recreation”. 

Measures 

Literature on retailing, services marketing and environmental psychology was used to 

identify pre-existing measures. Measures were then adapted to the online retailing context. Two 

studies were realized in order to adapt and validate our measures. 

The first study evaluated each scale performance: 60 students had to visit two electronic 

catalogs and to respond to a questionnaire after each visit. Online shopping hedonic or utilitarian 

values were assessed by the Personal Shopping Value (PSV) scale (Babin, Darden and Griffin 

1994). Four items measure hedonic value and three items measure utilitarian value. Responses are 

anchored on a 7 point scale from 1=strongly disagree to 7= strongly agree. Coefficient alpha is 

0.8767 for online shopping hedonic value and 0.8411 for online shopping utilitarian value. 

Satisfaction is assessed by a scale by Llosa (1996). One item measure the cognitive dimension of 

63

satisfaction (Oliver, 1980; Westbrook and Oliver, 1981; Oliver and Bearden, 1983), one item its 

affective dimension (Oliver, 1980; Westbrook, 1987; Westbrook and Oliver, 1981; Oliver and 

Bearden, 1983; Oliver, 1996) and two items its conative dimension (Westbrook and Oliver, 1981) 

(Hausknecht, 1990). Responses are anchored on a 7 point scale. Coefficient alpha is 0.9305. 

A second study developed a measure of shopping orientation. Expectations were assessed on 

a seven point semantic differential scale by a single item: “I expect: to be efficient � to spend a 

good time”. Consumers’ responses to their environment have been classified according to four 

dimensions representing affective quality: (1) arousing versus sleepy, (2) exciting versus gloomy, 

(3) pleasant versus unpleasant and (4) distressing versus relaxing (Russell, 1980; Russell and Pratt, 

1980; Darden and Babin, 1994; Babin and Darden, 1996). In this research, we use the pleasant 

versus unpleasant dimension for a parsimonious representation of affective quality. Eleven items 

measure consumer perceptions of the web site’s affective quality and were pre-tested with 110 

students. Five items were selected. Responses were anchored on a 7 point scale from 1= strongly 

disagree to 7=strongly agree. Coefficient alpha is 0.9114. 

Shopping orientation manipulation 

Scenarios have been used by prior research (Eroglu and Machleit, 1990; Dabholkar, 1994; 

Dabholkar and Bagozzi, 2002) to manipulate the shopping or service situation. In this research, two 

scenarios were created to induce a “shopping as work” and a “shopping as recreation” orientation. 

We used consumer role, time pressure and shopping motives to effect the manipulation and we 

controlled the effects of the physical and social environment - each subject visited the electronic 

catalog in a similar setting- computer laboratories of business schools – and antecedent states – each 

subject had to visit an electronic catalog with a “shopping as work” or a “shopping as recreation” 

scenario. A scenario was developed for each electronic catalog for a total of 8 scenarios. 

64

In the “shopping as work” condition, students had five minutes to visit an online travel (time 

pressure) agency web site to plan a specific trip – a destination, hotel description, room description 

and travel conditions were provided – (prepurchase information search) for their parents (shopping 

for somebody else) and send them an email. In the “shopping as recreation” condition, students had 

unlimited time (no time pressure) to visit an online travel agency to find ideas – only some general 

ideas about possible destinations were provided - (browsing) for their next trip or holiday (shopping 

for themselves). 

A further study was conducted using 58 students to pre-test each scenario. Our goal was to 

test the existence of significant differences between the “shopping as recreation” and “shopping as 

work” scenario with regards to affective quality and expectations. Students were instructed to read 

the scenario, to put themselves in the particular situation described by the scenario and then to 

complete a questionnaire measuring the affective quality of the online shopping situation described 

in the scenario and their expectations in this situation. Each student had to read a “shopping as 

work” and “a shopping as recreation” scenario. A paired sample t-test show that the affective 

quality of the online shopping situation is significantly higher for the “shopping as recreation” 

scenario (t = -11.274, p=0.000) and that consumers expects recreation to a greater degree for the 

“shopping as recreation” scenario (t = -11.634, p = 0.000). This difference was validated for each 

electronic catalog. 

A manipulation check was introduced in the final questionnaire: “How did you consider 

the visit before visiting the web site: as work � as recreation”. 

Data analysis 

A technique proposed by Sauer and Dick (1993) and applied by Homburg and Giering (2001) 

was used to test moderating variables in a structural equation model (Hypotheses H1, H2, H5 and 

65

H6). It involves performing a multi-group analysis with each sub-sample of the moderating 

variables (gender and shopping orientation). Two models are compared: one model in which the 

parameters are allowed to vary freely across groups (general model) and another model in which 

parameters are constrained to be equal across groups (restricted model). If there is a significant 

difference between the groups, the general model should fit better (χ²) than the restricted model. It is 

necessary to calculate the difference between the χ² of the general model and the χ² of the restricted 

model (dχ²). A significant dχ² suggest that the relationships are different between the groups and is 

consistent with moderation. 

Research results 

Confirmatory factor analysis (CFA) was conducted on all measures: online shopping values 

and consumer satisfaction with the electronic catalog visit. Two items had high modification indices 

and were dropped from the analysis. Another CFA was conducted. Fit indices are correct for each 

scenario: χ²=42.4. RMSEA=0.049, RMR=0.078, GFI=0.972, AGFI=0.947, TLI=0.985, CFI=0.990 

for the “shopping as recreation” scenario and χ²=38.5, RMSEA=0.043, RMR=0.109, GFI=0.974, 

AGFI=0.950, TLI=0.988, CFI=0.992 for the “shopping as work” scenario. Correlations between 

constructs were examined to test discriminant validity. Some variables such as satisfaction and 

online shopping hedonic value are highly correlated in the “shopping as recreation” condition (see 

table 1). 

Table 1 – Correlation between variables in the “shopping as recreation” scenario 

Satisfaction Hedonic value Utilitarian value 

Satisfaction 1 0.828 0.660 

Hedonic value 0.828 1 0.454 

Utilitarian value 0.660 0.454 1 

66

This pattern is consistent with our research hypotheses. Online shopping hedonic value was 

predicted to have a relatively stronger effect on consumer satisfaction in this condition. A procedure 

used by Dabholkar and Bagozzi (2002) which involves conducting a simple CFA with the two 

constructs as one was applied in this scenario to assess the degree of discriminability in light of the 

high correlations. In each case, the CFA with two constructs has better fit indices than the CFA with 

two constructs as one. Coefficient alphas are correct for each scale: 0.832 for online shopping 

hedonic value, 0.878 for online shopping utilitarian value and 0.907 for consumer satisfaction. 

Manipulation check 

A single item was used to assess manipulation success: : “How did you consider the visit before 

visiting the web site: as work � as recreation”. A paired sample t-test was realized to compare consumer 

responses to this question for each scenario. Results show a significant difference (t=4.714, p=0.000) 

between the “shopping as recreation” scenario (4.608) and for the “shopping as work” scenario (4.232). 

Moderating impact of shopping orientation 

Model fit is satisfying: χ²=80.862, RMSEA=0.033, RMR=0.095, GFI=0.973, AGFI=0.949, 

TLI=0.987, CFI=0.991. In the “shopping as recreation” condition (see figure 3), consumer 

satisfaction with the electronic catalog visit is determined mainly by online shopping hedonic value 

(0.666). The correlation between online shopping hedonic value and online shopping utilitarian 

value is 0.454. In the “shopping as work” condition (see figure 3), consumer satisfaction with the 

electronic catalog is determined by both visit hedonic (0.473) and utilitarian (0.469) value. The 

correlation between online shopping hedonic value and online shopping utilitarian value is 0.266 

67

Differences between the “shopping as recreation” and the “shopping as work” models were 

tested globally and at the individual parameter level. These analyses suggest: (1) a significant 

difference in overall fit between the models (dχ² =21.661 dDL=9, p

Figure 3 – Shopping orientation and the value-satisfaction chain 

Hypotheses 3 and 4 are supported. Online shopping hedonic value is significantly higher in 

the “shopping as recreation” scenario (-4.743, p

(Hypothesis 5) and online shopping utilitarian value has a stronger impact on consumer satisfaction 

for males, but the difference is not significant (Hypothesis 6). 

Hypothesis 7 is supported. Online shopping hedonic value is significantly higher for females 

than for males (2.621, p

Should an online shopping web site deliver both hedonic and utilitarian value? Both online 

shopping hedonic and utilitarian shopping value have a strong impact on consumer satisfaction. Our 

results suggest however that consumers attach importance to the hedonic aspects of a web site visit, 

whatever the orientation. The online shopping hedonic value has a stronger impact on consumer 

satisfaction when shopping for recreation than when shopping with a specific task in mind. 

However, the online shopping utilitarian value does not have a stronger impact on consumer 

satisfaction when shopping with a specific task in mind than when shopping for recreation. 

Consistent with patronage theory, online shopping hedonic value is higher when shopping for 

recreation and utilitarian value is higher when shopping with a specific task in mind. Previously, 

this hypothesis was not supported (Hammond, McWilliam and Diaz, 1998). Our results also show 

differences between males and females evaluation of the e-shopping experience. Although the 

differences are not significant, online shopping utilitarian value has a stronger impact on consumer 

satisfaction for males and online shopping hedonic value has a stronger impact on consumer 

satisfaction for females. Females experience greater hedonic value than males when shopping on 

the Internet. Interestingly, the link between hedonic value and utilitarian value is higher for females 

than for males which mean shopping efficiency and shopping enjoyment are linked to a greater 

extent for females. 

Our findings suggest that designers should consider both hedonic and utilitarian aspects of a 

web site, but should take into account the consumer’s orientation and gender. We showed that 

females experienced greater hedonic value when shopping on the Internet and that shopping 

efficiency and shopping enjoyment were closely related for them. Consumers evaluate online 

shopping mainly on its hedonic aspects when shopping for hedonic purposes and on both hedonic 

and utilitarian aspects when shopping with a specific task orientation. This research also 

71

demonstrates the importance of measuring both hedonic and utilitarian value and of controlling for 

shopping orientation and gender when pre-testing an electronic catalog. 

Some limitations of our study should be noted. First, our study only considers electronic 

catalogs from online travel agencies. As such the external validity may be limited. Generalizability 

could be established if additional studies validate the conceptual framework for other industries 

such as banking or computers. Second, the conceptual framework was tested with students. 

Although according to Calder, Phillips and Tybout (1981), using students reduces extraneous 

sources of variance, additional research is needed to validate the results with a non student sample. 

Third, subjects who had already visited the electronic catalog were excluded from the final sample. 

This procedure was necessary to ensure the internal validity of our research but limits its external 

validity. 

References 

Arnold, Mark J. Reynolds Kristy E. Hedonic shopping motivations. Journal of Retailing 2003; 79, 

2: 77-95 

Babin Barry J. Attaway Jill S. Atmospheric affect as a tool for creating value and gaining share of 

customer. Journal of Business Research 2000; 49: 91-99. 

Babin Barry J. Darden William R. Good and bad shopping vibes: spending and patronage 

satisfaction. Journal of Business Research 1996; 35 (3): 201-206. 

Babin Barry J. Darden William R. Griffin Mitch. Work and/or fun: measuring hedonic and 

utilitarian shopping value. Journal of Consumer Research 1994; 20: 644-656. 

Babin Barry J. The in-store retail experience: A CEV approach to consumer shopping activity. 

Dissertation, Louisiana State University, 1991. 

72

Belk Russell W. Situational variables and consumer behavior. Journal of Consumer Research 1975; 

2 (3): 157-164. 

Bellenger, Danny N. Korgaonkar P.K. Profiling the recreational shopper. Journal of Retailing 

1980; 56 (3): 77-92 

Bellenger, Danny, Robertson Dan H. Greenberg Barnett A. Shopping center patronage. Journal of 

Retailing 1977; 53 (2): 29-38 

Bird, Michael. Let’s Talk about Sex. In Store Marketing 2002; July 2002: 21-22 

Bloch Peter H. Ridgway Nancy M. Sherrell Dan L. Extending the concept of shopping: An 

investigation of browsing activity. Journal of the Academy of Marketing Science 1989; 17 (1): 13- 

21. 

Bloch Peter H. Sherrell Dan L. Ridgway Nancy M. Consumer search: An extended framework. 

Journal of Consumer Research 1986; 13: 119- 126. 

Calder Bobby J. Phillips Lynn W. Tybout Alice M. Designing research for application. Journal of 

Consumer Research 1981; 8 (2): 197-207. 

Chen, Zhan Dubinsky Alan A conceptual model of perceived customer value in e-commerce: A 

preliminary investigation. Psychology and Marketing 2003; 20: 323-347. 

Childers Terry L. Carr Christopher L. Peck Joann Carson Stephen Hedonic and utilitarian 

motivations for online retail shopping behavior. Journal of Retailing 2001; 77: 511-535. 

Cooper-Martin, E. (1991), Consumers and movies: some findings on experiential products, R. 

Holman and M. Solomon (eds), Advances in Consumer Reseach, Provo UT: Association for 

Consumer Research, 372-378 

Dabholkar Pratibha A. Bagozzi Richard P. An attitudinal model of technology based self-service: 

moderating effects of consumer trait and situational factors. Journal of the Academy of Marketing 

Science 2002; 20 (3): 184-201. 

73

Dabholkar, Pratibha A. Incorporating choice into an attitudinal framework: analyzing models of 

mental comparison processes. Journal of Consumer Research 1994; 21: 100-118 

Darden William R. Babin Barry J. Exploring the concept of affective quality: expanding the concept 

of retail personality. Journal of Business Research 1994; 29 (2): 101-109. 

Dawson Scott Bloch Peter H. Ridgway Nancy M. Shopping motives, emotional states, and retail 

outcomes. Journal of Retailing 1990; 66 ( 4): 408-427. 

De Ruyter Ko Wetzels Martin Lemmink Jos Mattsson J. The dynamics of the service delivery 

process: a value-based approach. International Journal of Research in Marketing. 1997; 14: 231- 

243. 

Donovan, R. J. Rossiter John R. Store atmosphere: an environmental psychology approach. Journal 

of Retailing, 1982; 58 (Spring): 34-57. 

Eroglu, Sevgin A. Machleit, Karen A. An empirical study of retail crowding: antecedent and 

consequences. Journal of Retailing 1990; 66 (2): 201-221 

Ezell Hazel F Motes William H. Differentiating between the sexes: a focus on male-female grocery 

shopping attitudes and behavior. Journal of Consumer Marketing 1985; 2 (2); 29-40. 

G.V.U. GVU’s 10 th WWW User Surveys, http://www.gvu.gatech.edu/ 1999 

Griffin Mitch Babin Barry J. Modianos Doan Shopping value of russian consumers: the impact of 

habituation in a developing economy. Journal of Retailing 2000; 76 (1): 33-52. 

Hammond Kathy McWilliam Gil Diaz Andrea Fun and work on the web: differences in attitudes 

between novices and experienced users. Advances in Consumer Research. J.W. Alba J.W. 

Hutchinson (eds). Provo UT: Association for Consumer Research (25): 372-378. 

Hausknecht, Douglas R. Measurement scales in consumer satisfaction/dissatisfaction. Journal of 

Consumer Satisfaction, Dissatisfaction and Complaining Behavior 1990; 3: 1-11 

Hirschman, Elisabeth C. Holbrook Morris B. Hedonic consumption: emerging concepts, methods 

and propositions. Journal of Marketing 1982; 46 (3): 92-101 

74

Hoffman Donna L. Novak Thomas P. Marketing in hypermedia computer-mediated environments: 

conceptual foundations. Journal of Marketing 1996; 60 (3): 64-77. 

Holbrook, M.B. The Nature of Consumer Value: An Axiology of Services in the Consumption 

Experience. In Richard T. Rust and Richard T. Oliver, editors, Service Quality: New Directions in 

Theory and Practice., Sage Publications Inc.: Thousand Oaks, California, 1994, pp. 21-71 

Holbrook, Morris B. Hirschman, Elisabeth C. The experiential aspects of consumption: consumer 

fantasies, feelings and fun. Journal of Consumer Research 1982; 9 (2): 132-140 

Holbrook, Morris B. Consumer value: a framework for analysis and research. Routledge 

Intepretative Marketing Research Series: London and New York, 1999 

Holbrook, Morris B. Emotion in the consumption experience: toward a new model of the human 

consumer, In R.A. Peterson, H.D. Hoyer and W.R., Wilson, editors, The Role of Affect in 

Consumer Behavior: Emerging Theories and Applications¸ Lexington Books, D.C. Heath and 

Company/Lexington, Massachusetts/Toronto, 1986, pp. 17-52 

Homburg Christian Giering Annette Personal characteristics as moderators of the relationship 

between customer satisfaction and loyalty – an empirical analysis. Psychology and Marketing 2001; 

18 (1): 43-66. 

Journal du Net Les Chiffres Clés, http://www.journaldunet.com/chiffres-cles.shtml 2003 

Lemmink Jos de Ruyter Ko Wetzels Martin The role of value in the delivery process of hospitality 

services. Journal of Economic Psychology 1998; 19: 159-177. 

Llosa, Sylvie Contributions à l’Etude de la Satisfaction dans les Services, Dissertation, Université 

de Droit, d’Economie et des Sciences d’Aix Marseille, Institut d’Administration des enterprises, 

2001 

Lutz Richard J. Kakkar Pradeep Situational influence in interpersonal persuasion. Advances in 

Consumer Research, Beverlee Anderson (ed). Ann Arbor, MI: Association for Consumer Research. 

1976; 3: 370-378. 

75

Mano Haim Elliott Michael T. Smart shopping: the origins and consequences of price savings. 

Advances in Consumer Research Merrie Brucks and Deborah J. MacInnis (Eds.) Provo UT: 

Association for Consumer Research. 1997; (24) 504-510. 

Mano Haim Oliver Richard L. Assessing the dimensionality and structure of the consumption 

experience: evaluation, feeling, and satisfaction.” Journal of Consumer Research 1993; 20: 451- 

466. 

Mao Wendy W. Fader Peter S. Dynamic conversion behavior and e-commerce sites. Working 

Paper. Wharton Business School 2002. 

Mathwick Charla Malhotra Naresh K. Rigdon Edward Experiential value: conceptualization, 

measurement and application in the catalog and Internet shopping environment. Journal of Retailing 

2001; 77: 39-56. 

Oliver Richard L. Measurement and evaluation of satisfaction processes in retail settings. Journal of 

Retailing 1981; 57 (3): 25-48. 

Oliver, Richard L. A cognitive model of the antecedents and consequences of satisfaction decisions. 

Journal of Marketing Research 1980; 27: 460-469 

Oliver, Richard L. Bearden, William O. The role of involvement in satisfaction processes. In 

Richard Bagozzi and Ann Tybout, editors, Advances in Consumer Research, 10, Ann Arbor, MI: 

Association for Consumer Research, 1983: pp. 250-255 

Otnes Cele McGrath Mary Ann. Perceptions and realities of male shopping behavior. Journal of 

Retailing 2001; 77 (1): 111-137. 

Raman Niranjan. V. Leckenby John D. Factors affecting consumer « WebAd » visits. European 

Journal of Marketing 1998; 32 (7/8): 737-748. 

Russell James A. A circumplex model of affect. Journal of Personality and Social Psychology 1980; 

39 (6): 1161-1178. 

76

Russell James A. Affective space is bipolar. Journal of Personality and Social Psychology 1980; 37 

(3): 345 – 356. 

Russell James A. Pratt G. A description of the affective quality attributed to environments. Journal 

of Personality and Social Psychology 1980; 38 (3): 311 – 322. 

Sauer Peter L. Dick Alan. Using moderator variables in structural equation models. Advances in 

Consumer Research Leight McAlister and Michael Rothschild (eds). Provo UT: Association for 

Consumer Research 1993; (20): 636-640. 

Seiders, Kathleen, Leonard Berry Gresham, Larry G. Attention Retailers! How Convenient is your 

Convenience Strategy. Sloan Management Review 2000: 79-89 

Senecal Sylvain Gharbi Jamel E. Nantel Jacques The influence of flow on hedonic and utilitarian 

shopping values. Advances in Consumer Research. S Broniarczyk and K. Nakamoto (eds) Provo, 

UT: Association for Consumer Research 2001. (29), www.acrweb.com. 

Szymanski David M. Hise Richard T. E-Satisfaction: An initial examination. Journal of Retailing 

76 2000; (3): 309-322. 

Szymanski, David M. Hise Richard T. E-satisfaction: an initial examination. Journal of Retailing 

2001; 76: 309-322. 

Tauber Edward M. Why do people shop? J.ournal of Marketing 1972; 36: 46-49. 

Venkatesh Viswanath Morris Michael G. Why don’t men ever stop to ask for Directions? gender, 

social influence, and their role in technology acceptance and usage behavior. MIS Quaterly 2000; 

24 (1): 115-139. 

Westbrook Richard A. Black William C. A motivation-based shopper typology. Journal of Retailing 

1985; 61 (1): 78-103. 

Zeithaml V., Parasuraman A. & Malhotra Arvind A conceptual framework for understanding e- 

service quality: implications for future research and managerial practice. Report N°.00.115. 

Marketing Science Institute. Working Paper Series 2000 

77

Zeithaml V., Parasuraman A. Malhotra Arvind Service quality delivery through Web sites: A 

critical review of extant knowledge. Journal of the Academy of Marketing Science 2002; 30: 162- 

175. 

78

The Influence of Intangibility on Perceived Evaluation Difficulty and Risk: Brand and 

Generic Product-Category Perspectives 

Michel Laroche, Concordia University 

Maria Kalamas, Concordia University 

Gordon H. G. McDougall, Wilfrid Laurier University 

Filip Bartos, Concordia University 

Yi Zhong, Concordia University 

Abstract 

As virtual products/services have emerged in consumer markets, intangibility has assumed an important role 

in marketing, shifting from a service-exclusive term to a product-related one as well. A three-dimensional 

intangibility construct, which includes physical intangibility, mental intangibility, and generality, has 

recently been identified. While this finding has advanced our understanding of intangibility per se, little has 

been done in relating this new operationalization to consumer behavior and especially branding strategies. 

Thus, the focus of this research is to study whether brands, as opposed to generic product-categories, 

efficiently reduce risk and evaluation difficulty when the product/service is perceived to be intangible. Our 

results show brands to be major intangibility-reducers for services but not for products. Brands also variably 

reduce the effects of the three types of intangibility on perceived evaluation difficulty and risk. In our holistic 

model, knowledge and/or involvement moderate the relationships between intangibility and, perceived risk 

and/or difficulty of evaluation. We also uncover differences between the moderating effects of knowledge 

and involvement in our comparisons of the branded and generic contexts. In light of our findings, we make 

brand-related recommendations, which are of interest to both academicians and practitioners. 

79

1. Introduction 

Like inseparability, heterogeneity, and perishability, intangibility is one of the four major 

characteristics used to classify services (Rust et al. 1996; Shostack 1977; Zeithaml et al. 1985). 

Considered to be the key characteristic that differentiates products from services (Bebko 2000; 

Zeithaml and Bitner 2000), the role of intangibility has shifted from a service-exclusive term to a 

product-related one as well. Faced with the changes and challenges brought forth by a new 

category of virtual or semi-virtual products/services (e.g., Web browsers, MP3 music files), 

academicians are now equipped to study the effects of intangibility on perception of risk and 

evaluation difficulty using a three-dimensional intangibility construct. Developed by Laroche, 

Bergeron and Goutaland (2001), the intangibility construct now comprises three dimensions: 

physical intangibility, mental intangibility, and generality. Each of these dimensions may have a 

differential impact on consumers’ decision making processes both for branded and generic 

products/services. The importance of studying intangibility within a branding context rests on 

the premise that brands are believed to reduce consumers’ perceived risk and difficulty of 

evaluation. 

Given the dearth of research on intangibility, the focus of this study is, first, to confirm 

the with new data the three dimensions of intangibility (Laroche, Bergeron and Goutaland 2001) 

and then to study the effects of intangibility on perceived risk and difficulty of evaluation (for 

the proposed model, see Figure 1). Another aim of this study is to determine how knowledge and 

involvement, acting as moderators, affect the relationships between intangibility and difficulty of 

evaluation, and intangibility and perceived risk. By comparing generic products/services with 

branded ones, we also explore how the two strategies impact the relationships among the 

proposed constructs. The overall goal of this research is thus to determine whether branding 

80

educes risk and evaluation difficulty when the branded product/service is perceived as 

intangible. 

2. Literature Review and Hypotheses 

2.1. Intangibility 

2.1.1. Definition of Intangibility 

[Figure 1 about here] 

Lovelock (2001) recently identified nine basic differences between products and services 

with intangibility as one of the major distinguishing factors. Shostack (1977) illustrated the 

differences between tangible products and intangible services by suggesting that the former 

could be described precisely, be replicated exactly, be modified, and be duplicated while the 

latter were dynamic, evolving, and hard to quantify. Along the same lines, Berry (1980) defined 

intangibility as something impalpable. Adding another level of complexity, Bebko (2000) 

recently proposed that intangibility should include lack of physical evidence of the process 

instead of being defined exclusively as lack of physical attributes of the outcome. 

2.1.2. Dimensions of Intangibility 

Usually defined as inaccessible by the senses, subjective, and difficult to measure, most 

scholars agree that intangibility is one of the most distinguishing characteristics of services 

(Bateson 1979, Berry 1980, Lovelock 2001, Shostack 1977, Van Dierdonck 1992). 

Characterizing services as physically and mentally intangible, a number of researchers 

conceptualized intangibility as a two-dimensional construct (Bateson 1979; Berry 1980; 

McDougall and Snetsinger 1990). Among them, Dubé-Rioux, Regan and Schmitt (1990) 

proposed a two-dimensional intangibility construct composed of concreteness and specificity. 

Their findings revealed that services differed in their cognitive representations with abstract 

services represented by means of generic attributes and concrete services represented using more 

81

specific attributes. Dealing with products, Breivik, Troye and Olsson (1998) also developed a 

two-dimensional model of intangibility with inaccessibility to the senses and generality as the 

two components. Compared to tangible product attributes, which are perceived upon exposure, 

intangible product attributes are mentally constructed and hence, subject-dependent and 

inaccessible to the senses. Generality, in this study, encompassed the general product attributes 

that lead to general outcomes (e.g., car safety). Common among all the studies described thus far 

is a two-dimensional conceptualization of intangibility. 

Unlike previous conceptualizations, the most recent study proposed a three-dimensional 

construct of intangibility, composed of physical intangibility (i.e., not easy to see or touch), 

mental intangibility (i.e., not easy to grasp mentally), and generality (as opposed to specificity) 

(Laroche, Bergeron and Goutaland2001). The finding of the third dimension, mental 

intangibility, confirmed Laroche, Bergeron and Goutaland (2001)’s proposition that 

inaccessibility to the senses (i.e., physical intangibility) and generality were not sufficient to 

fully measure the overall construct since intangibility was subject-specific (Hirschman 1981, 

Breivik, Troye and Olsson 1998). Briefly, Laroche, Bergeron and Goutaland (2001) concluded 

that the new three-dimensional concept of intangibility provides valuable insights into the 

tangibility of goods and the intangibility of services. In all, their findings showed that goods 

could be both physically tangible and mentally intangible, whereas services could be both 

physically intangible and mentally tangible. 

2.1.4. Intangibility and Related Marketing Strategies 

As a result of its inherent characteristics, intangibility has brought forth many challenges 

for marketers including storability (Berry 1980), communicability (Rathmell 1974), risk 

perceptions, and evaluation difficulties. To deal with the inherent problems of intangibility, 

academicians recommend tangibilizing products/services through management of physical 

82

surroundings and presentations (Berry and Clark 1986), mental visualization and association 

(Berry and Clark 1986), effective advertising (Berry and Clark 1986, Mittal and Baker 2002, 

Zeithaml and Bitner 2000), and the use of image and brand management (Berry 2000, Edgett and 

Parkinson 1993). 

2.1.5. Intangibility from the Brand and Generic Product-Category Perspectives 

Consumer choices follow a somewhat hierarchical process whereby different choices are 

related to different levels of abstraction (Howard 1977). Choices at the product-category level 

are higher-level, more abstract choices, while lower-level choices involve more concrete 

alternatives such as the brands within a category. Unlike Howard’s hierarchical process, Johnson 

and Fornell (1984) hypothesized a continuum of attributes from concrete to abstract and showed 

that the more noncomparable (dissimilar) products become the more abstract their product 

attributes. Building on these hypotheses, the authors showed that specific product attributes are 

associated with brands while generic attributes are associated with product categories. Since 

generic product categories are more abstract and brands are more specific and concrete, we 

hypothesize that: 

H1: The degree of generality associated with generic category-level choices is higher than it is for brand-level choices. 

Mental tangibility and specificity appear to be the most efficient tools for reducing customer perceived intangibility especially for 

information and virtual products (Laroche, Bergeron and Goutaland 2001). Along the same lines, Mittal and Baker (2002) suggested 

that brand identity and brand positioning were useful advertising strategies in fighting against the intangibility of hospitality 

services. Brands, like pictures and company logos, may help consumers visualize services and therefore, increase mental tangibility. 

Based on the limited evidence in the services literature, we infer that brands reduce mental intangibility for both products and 

services, leading us to hypothesize that: 

H2: The degree of mental intangibility associated with generic category-level choices is higher than it is for brand-level choices. 

Since physically intangible products are not easily seen or touched, physical intangibility should be independent of whether the 

product is branded or belongs to a generic product category. Be it a Sony or a generic version, for example, the physical intangibility 

of an MP3 file always exists. For services, however, the independence of physical intangibility from brand and generic perspectives is 

questionable. Being physically intangible, services necessitate a physical representation strategy aimed at enhancing service tangibility 

through physically accessible objects, which are directly or peripherally a component of the service (e.g., buildings) (Berry and Clark 

1986). For instance, some hotels purposely park luxury cars before their entrances in order to make their services appear more 

physically tangible (Kotler et al. 1999). We thus infer that the physical representation required to tangibilize services will be more 

easily achieved through branding as opposed to a generic strategy. 

H3: The degree of physical intangibility associated with generic category-level choices is a) higher than it is for brand-level choices 

of services and b) similar to it for brand-level choices of products. 

83

2.2. Perceived Difficulty of Evaluation 

Intangibility makes it harder for marketers to promote intangible products/services and for 

consumers to evaluate offerings with intangible attributes (Zeithaml 1981). Much of the 

literature in services marketing has linked intangibility to consumers’ decision-making 

difficulties including decreased certainty of evaluation (Mitchell and Greatorex 1993; Murray 

1991) and increased difficulty of evaluation (McDougall 1987; McDougall and Snetsinger 1990; 

Zeithaml 1981) and processing effort (McDougall 1987). Zeithaml postulated that most goods 

fell to the left of the tangibility-intangibility continuum because most goods were easy to 

evaluate, whereas most services fell to the right since most services were difficulty to evaluate. 

Empirically corroborating Zeithaml’s postulations, Goutaland (1999) found mental intangibility 

and generality to be positively related to difficulty of evaluation and no relationship between 

physical intangibility and difficulty of evaluation. Likewise, Breivik, Troye and Olsson (1998) 

found a positive association between generality and difficulty of evaluation but contrary to 

expectations, their empirical study also revealed that inaccessibility to the senses was negatively 

associated with perceived difficulty of evaluation. 

2.3. Perceived Risk 

2.3.1. Types of Perceived Risk 

Cox and Rich (1964) first identified two types of perceived risk namely, financial risk 

and social-psychological risk. Later on, Jacoby and Kaplan (1981) suggested that overall 

perceived risk should include performance risk, physical risk, social risk, psychological risk, and 

time risk. Extending previous conceptualizations, researchers identified six types of risk (i.e., 

financial risk, performance risk, physical risk, psychological risk, social risk and convenience 

loss or time-related risk) (Murry and Schlacter 1990; Stone and Grønhaug 1993). Although 

Murry and Schlacter (1990) believed that all six types of risks were perceived as greater for 

84

services than for goods, financial and performance risks were found not to be statistically 

significant. 

2.3.2. Risk Reduction 

Perceived risk is generally recognized as being influential on consumer decision-making 

because of the uncertainty and possible negative consequences involved with consumption. 

Bauer (1960) suggested that consumers might have preferences towards risk reduction strategies 

citing brand loyalty as a preferred risk reducer. Among the risk reduction strategies explored and 

recommended by researchers, branding is recognized as a major risk-reducer (Roselius 1971, 

Sheth and Venkatesan 1968, Wernerfelt 1988). Included in the arsenal of risk-reduction 

strategies we also find advertising (Barach 1969; Cox 1967), word of mouth (Cunningham 1967; 

Roselius 1971), company loyalty (Bauer 1961), personal and group influence (Perry and Hamm 

1968), and price-quality issues (Roselius 1971). 

2.3.3. Difficulty of Evaluation, Perceived Risk, and Intangibility 

Using a different conceptualization of intangibility, Breivik, Troye and Olsson (1998) 

found inaccessibility to the senses to be negatively related to difficulty of evaluation. McDougall 

drew a similar conclusion in an exploratory study where tangibility was conceptualized as “easy 

to picture or visualize before purchase” (1987, p. 430). He later explained that the nonsignificant 

influence of physical intangibility on difficulty of evaluation might have been due to the 

oversimplified tangibility measure and non-representative sample used in the research. Likewise, 

Goutaland (1999) found no significant relationship between the two constructs. Despite these 

findings, we posit a positive relation between physical intangibility and difficulty of evaluation. 

With respect to the other dimensions of intangibility, namely generality and mental intangibility, 

no specific studies exist from which we can draw, thus leading us to infer that the more general 

85

and mentally intangible a product/service is the more difficult it will be for consumers to 

evaluate it. 

Previous research suggests the more intangible a product/service is, the greater the 

perceived risk (Mitchell and Greatorex 1993; Murray and Schlacter 1990; Zeithaml 1981). While 

empirically demonstrated, this finding is solely based on the physical intangibility construct. In 

Goutaland’s study (1999) mental intangibility was positively related to perceived risk while 

physical intangibility and generality were not. Compared to mental intangibility, physical 

intangibility and generality appeared to contribute very little in stimulating perceived risk, 

especially in the context of a three-dimensional intangibility construct. While past literature 

supports the idea that intangibility results in greater risk perceptions, the links between the 

multidimensional intangibility and risk constructs necessitate further exploration. Based on 

previous findings, the following hypotheses ensue: 

H4: The more mentally intangible a product/service is, a) the more difficult it is to evaluate 

and b) the more risky it is perceived. 

H5: The more general a product/service is, a) the more difficult it is to evaluate and b) the 

more risky it is perceived. 

H6: The more physically intangible a product/service is, a) the more difficult it is to evaluate 


2.3.4. Difficulty of Evaluation and Perceived Risk at the Brand and Product-Category Levels 

Brands facilitate consumer decision making particularly by improving decision-making 

efficiency (Alba and Hutchinson 1987; Johnson and Russo 1984) and reducing perceived risk 

(Wernerfelt 1988). Umbrella branding also eliminates consumers’ perceived risks, difficulty of 

evaluation, and product uncertainty (Erdem 1998). Branding plays an especially important role 

in the service arena by enabling customers to tangibilize and visualize the service and thereby 

reduce perceived social and monetary risks (Berry 2000). While Berry did not empirically test 

86

his propositions, his arguments on branding, perceived risk, and tangibilization are constructive. 

Extending previous findings to incorporate intangibility, we hypothesize that: 

H7: The impact of intangibility on a) difficulty of evaluation and b) perceived risk is greater 

for generic-category level choice than brand-level choice. 

2.4. Knowledge 

As a unidimensional construct (Alba and Hutchinson 1987), knowledge was traditionally operationalized 

using such measures as frequency of purchase (Anderson et al. 1979; Bettman and Park 1980), objective tests 

(Brucks 1985), and self-report measures (Johnson and Russo 1984). Given the diversity and difference in measures 

of knowledge and the inherent complexity of the construct, researchers called for a multi-dimensional 

operationalization (Bettman 1986, Brucks 1986; Brucks and Mitchell 1981). Some proposed a two-dimensional 

measure of consumer knowledge with familiarity and expertise as the two factors (Alba and Hutchinson 1987), 

while others came up with experience and expertise as the two dimensions. Murry and Schlacter (1990) proposed 

five items to measure experience (i.e., purchase experience, product utilization and exposure, brand familiarity, 

purchase frequency, and purchase confidence) and concluded that experience reduces consumers’ perceived risk. 

The most common approach used by researchers to operationalize experience is to treat it as familiarity (Alba and 

Hutchinson 1987; Johnson and Russo 1984). Familiarity improves consumers’ ability to categorize information at a 

more specific rather than generic level (Alba and Hutchinson 1987). Expertise, the second dimension of knowledge 

and distinct from experience, is commonly defined as consumers’ capacity to successfully perform a product-related 

task (Alba and Hutchinson 1987). The authors proposed that expertise was composed of five dimensions: cognitive 

effort and automaticity, cognitive structure, analysis, elaboration, and memory. However, the authors did not 

provide operational methods to measure the proposed dimensions. Zaichkowsky (1985) suggested that expertise and 

experience were strongly related when expertise was measured subjectively, but the relationship was weak when it 

was measured objectively thus making it is necessary to examine both subjective and objective perceptions of 

knowledge. 

87

2.4.1. Knowledge and Intangibility 

In Goutaland’s empirical study (1999), subjective knowledge had a negative influence on 

both difficulty of evaluation and perceived risk. As a moderator, subjective knowledge interacted 

with physical intangibility (mental intangibility) to increase (decrease) perceived risk. However, 

the expected moderating effect of knowledge on the relationship between intangibility and 

difficulty of evaluation was not found. These somewhat inconsistent findings necessitate further 

testing of the relationships between knowledge, intangibility, risk, and difficulty of evaluation. 

Moreover, the moderating effect of knowledge will be retested from the brand and generic 

category perspectives. Since brands reduce perceived risk and difficulty of evaluation, as 

explained in earlier sections, the moderating effect of knowledge will be less in a brand context 

than in a generic one. We thus hypothesize that: 

H8: Knowledge moderates the relationships between the three dimensions of intangibility 

(i.e., generality, physical intangibility and mentally intangibility) and a) difficulty of 

evaluation and b) perceived risk. 

H9: The more knowledgeable a consumer perceives himself to be regarding a 

product/service a) the less difficult it is to evaluate it and b) the less risky it is perceived. 

H10: The moderating effects of knowledge on a) difficulty of evaluation and b) perceived 

risk at the generic level are greater than those at the brand level. 

2.5. Involvement 

Zaichkowsky (1985) stated that involvement had been diversely defined and measured 

due to its different applications. In fact, involvement, applied under various objectives, led to 

different responses for products (Howard and Sheth 1969), for advertisements (Krugman 1977), 

and for purchase decisions (Clarke and Belk 1978). For the purpose of developing a scale, 

Zaichkowsky defined involvement as “a person’s perceived relevance of the object based on 

inherent needs, values, and interests” (1985, p. 342). Based on the definition proposed, 

Zaichkowsky (1985) empirically developed the personal involvement inventory – a scale used to 

measure one’s product involvement. Other researchers argued that the construct was 

multidimensional rather than unidimensional. According to Laurent and Kapferer (1985), there 

88

was more than one type of involvement in consumer research as different antecedents of 

involvement resulted in different corresponding behaviors. As such, an involvement profile 

might serve as a better measurement in specifying the relationship between consumers and 

product categories. 

2.5.1. Involvement and Intangibility 

McDougall (1987) found that both intangibility and involvement were significantly 

related to product evaluation. Adding further insights, Goutaland’s empirical study (1999) 

revealed that involvement was positively associated with difficulty of evaluation and perceived 

risk. Both mental intangibility and generality, two of the three dimensions of intangibility, 

interacted with involvement to produce a negative effect on difficulty of evaluation. Moreover, 

the interaction between involvement and physical intangibility (generality) had a negative 

(positive) effect on perceived risk. Based on previous findings, both the predictive and 

moderating roles of involvement will be tested here. 

H11: Involvement moderates the relationships between the three dimensions of intangibility 

(i.e., generality, physical intangibility, and mental intangibility) and a) difficulty of 

evaluation and b) perceived risk. 

H12: The more involving a product/service is for a consumer, a) the more difficult it is to evaluate it 


H13: The moderating effects of involvement on a) difficulty of evaluation and b) perceived 

risk at the generic level are greater than those at the brand level. 

3. Methodology 

3.1. Questionnaire 

3.1.1. Product and Service Selection 

Selected based on their suitability for a student population, the goods (i.e., a pair of jeans, a 

computer, and a music CD) and services (i.e., a pizzeria dinner, a checking account, and an 

Internet browser) used in this study varied in their degree of intangibility. In each good/service 

89

category, we included both generic and well-known brands (i.e., a pair of Levi’s jeans, an IBM 

computer, a Beatles CD, Pizza Hut pizza, a Royal Bank checking account, and Netscape). 

3.1.2. Questionnaire Design 

We developed a total of sixteen versions of the questionnaire (Table 1), half of which related to 

online purchasing and the remainder focused on offline purchasing. In addition to demographic 

information such as age, gender, education, part-time or full-time student status, language, and cultural 

background, each questionnaire included measures of all constructs for one tangible good, one less 

tangible good, and one service. 

3.2. Measures 

3.2.1. Intangibility 

[Table 1 about here] 

We adopted the measures for the three dimensions of intangibility from Laroche, Bergeron and 

Goutaland (2001). 

3.2.2. Difficulty of Evaluation and Perceived Risk 

We measured difficulty of evaluation using this general statement: “Given that I have to acquire a 

product (“a product on the Internet” for online versions), choosing among the available brands will be” 

(“evaluating the product/service will be” for the online and brand versions) difficult, problematic, 

complex, and complicated. We measured perceived risk using five dimensions, namely, financial risk, 

psychological risk, performance risk, social risk, and time risk (Stone and Grønhaug 1993). 

3.2.4. Knowledge and Involvement 

In line with the literature, we measured both subjective knowledge and practical experience using 

items adapted from Biehal (1983), Oliver and Bearden (1983), and Park et al. (1994). Reducing 

Zaichkowsky’s (1985) 20-item measure to 11 items, based on Mittal (1989)’s arguments, Goutaland 

(1999) measured involvement as a unidimensional construct. We used 5 of the most relevant items from 

Goutaland (1999)’s 11-item measure of involvement. 

90

3.3. Pretests and Data Collection 

We pretested each version of the questionnaire with a small group of students, following which we 

made some minor modifications to the wording of the questions. During the pretests, we also verified 

participants’ level of familiarity with the chosen brands. The data collection took place at a major North 

American university where we randomly distributed a total of eight hundred questionnaires (i.e., fifty 

copies for each of the sixteen versions). 

4. Analysis and Results 

4.1. Sample 

We collected a total of 783 usable questionnaires (i.e., a 97.9% response rate), which were by- 

and-large evenly distributed among the sixteen versions, the online/offline offerings, and the 

brand/generic contexts (Table 2). Reflecting the student body, over half (59.1%) of the respondents were 

concentrated in the 21 to 25 age group and about a quarter were in the under 20 category. With only a 

small percentage of graduate students, the vast majority of respondents were undergraduates (93.7%). 

Approached during daytime classes, 84.3% of respondents were full-time students while the rest were 

part-time students. Males and females were evenly represented in our sample. 

4.2. Factor Analysis 

Before testing the hypotheses, we ran an exploratory factor analysis using principal component 

extraction and oblimin rotation, to verify the existence of the dimensions proposed in the literature review 

and to examine the reliability of the measures used. After deleting two items which cross-loaded on other 

factors, we reran the factor analysis and found the eleven distinct constructs originally posited (Table 3). 

Among them, time risk explained most of the variance (25.5%), followed by knowledge (16.4%), and 

involvement (9.1%). With Cronbach’s alphas exceeding 0.82, our measures are highly reliable. 

4.3. Test of Hypotheses 

4.3.1. T-tests: H1 to H3 

To conduct the following analyses across the brand and generic categories, the data file was rearranged 

such that each single good and service studied became a unit. This enlarged the sample size to 2,349 (783 

respondents*3 goods/services) responses for the twelve products and services studied. To test hypotheses 1 to 3, we 

then conducted t-tests and found H1a, H2a, and H3a to be supported (for a summary of all results, see Table 4). 

Whereas the branded goods exhibit higher levels of generality, and physical and mental intangibility, than the 

generic counterparts, the reverse holds true for services. We then compared the branded and generic versions of 

each good/service to determine differences in physical and mental intangibility, and generality. All the branded 

91

goods exhibit higher levels of physical intangibility than the generic ones. Across the board, the branded goods also 

display higher levels of mental intangibility and generality. Unlike the branded goods, the branded services display 

lower levels of generality, physical and mental intangibility. The one exception, the Royal Bank checking account, 

exhibits higher levels of generality than the generic option. 

[Tables 2 to 4 about here] 

4.3.2. Regression analysis: H4 to H13 

We used multiple regressions to test the remaining hypotheses, applying the stepwise method in 

order to minimize multicollinearity effects. A summary of the results follows: 

INTANGIBILITY, KNOWLEDGE, AND INVOLVEMENT WITH DIFFICULTY OF EVALUATION (TABLE 5). 

Supporting H4a, H5a, H6a, H9a, and H12a, the results showed the three dimensions of intangibility, 

knowledge, and involvement, to be significantly related to difficulty of evaluation. 

INTANGIBILITY, KNOWLEDGE, AND INVOLVEMENT WITH PERCEIVED RISK (TABLE 6). We found 

support for H4b in that mental intangibility was positively related to all five types of risk, but only partial 

support for H5b since generality was only related to social and psychological risk. Contrary to H6b, 

physical intangibility was negatively related to social risk. The hypothesized relationship between 

knowledge and perceived risk also held for all types of risk, except social risk, providing strong support 

for H9b. Likewise for H12b, which was partially supported by virtue of the positive relationship between 

involvement and social and financial risk. For time risk, the relationship was found to be negative, in the 

opposite direction of what was hypothesized. 

MODERATING EFFECTS OF KNOWLEDGE AND INVOLVEMENT: INTANGIBILITY AND DIFFICULTY 

OF EVALUATION (TABLE 7). While knowledge interacted with mental intangibility to reduce difficulty of 

evaluation, involvement did not interact with any of the three dimensions of intangibility. Hence, H8a 

was partially supported, whereas H11a was not. 

MODERATING EFFECTS OF KNOWLEDGE AND INVOLVEMENT: INTANGIBILITY AND PERCEIVED 

RISK (TABLE 8). Overall, we found partial support for H8b and H11b. Knowledge moderates the 

relationship between mental intangibility and social and psychological risk; generality and time risk; 

physical intangibility and social risk. Likewise, involvement moderates the relationship between mental 

intangibility and social and psychological risk but, the moderating effect follows a different pattern when 

92

it comes to generality and physical intangibility. There, involvement moderates the relationship between 

generality and time and social risk; physical intangibility and four types of risk, except time risk. 

INTANGIBILITY AND DIFFICULTY OF EVALUATION: BRAND AND GENERIC PRODUCT CATEGORIES 

(TABLE 9). Since generality and physical intangibility have a greater effect on difficulty of evaluation in 

the generic context, as opposed to the branded one, H7a is partially supported. 

INTANGIBILITY AND PERCEIVED RISK: BRAND AND GENERIC PRODUCT CATEGORIES (TABLE 10). 

Looking at the magnitude of the standardized coefficients, we find that physical intangibility has a greater 

effect on social, financial, and performance risk at the generic level as opposed to the branded one. For 

generality, the coefficients at the generic level are all greater than those at the brand level, with the pattern 

repeated for all five types of risk. For mental intangibility, the effect is opposite to what had been 

hypothesized in H7b, as it appears to have a greater impact at the brand level as opposed to the generic 

one, with the exception of social risk where the comparison between generic and brand contexts is 

consistent with the hypothesis. Based on these findings, H7b is partially supported. 

MODERATING EFFECTS OF KNOWLEDGE AND INVOLVEMENT ON DIFFICULTY OF EVALUATION: 

BRAND AND GENERIC PRODUCT CATEGORIES (TABLE 11). Although we hypothesized that the 

moderating effects of knowledge and involvement would be greater in the generic context, this only held 

true for the moderating effect of involvement on the relationship between physical intangibility and 

difficulty of evaluation. Hence, we found no support for H10a and only partial support for H13a. 

MODERATING EFFECTS OF KNOWLEDGE AND INVOLVEMENT ON PERCEIVED RISK: BRAND AND 

GENERIC PRODUCT CATEGORIES (TABLE 12). We found partial support for H10b and H13b in that the 

moderating effects of knowledge were evident between physical intangibility/generality and social risk; 

generality and psychological risk; and physical/mental intangibility and performance risk. Moreover, we 

found the hypothesized moderating effects of involvement between physical/mental intangibility and time 

risk; mental intangibility and social risk; mental/physical intangibility and psychological risk; and 

mental/physical intangibility and financial risk. 

5. Discussion 

[Tables 5 to 12 about here] 

93

5.1. Intangibility: Brand and Generic Product Categories (H1, H2, and H3) 

The proposed differences between the generic and brand perspectives for the three dimensions of 

intangibility (H1 to H3) held true for services. Physical and mental intangibility for both the checking 

account and the pizzeria dinner were lower for the branded services as opposed to the generic ones. While 

generality showed the same result for the pizzeria dinner, the Royal Bank checking account appeared as 

more general than the generic option. Perhaps the specific nature of the service (i.e., checking account), 

along with the online purchase scenario, reduce the impact of the brand on consumers’ perceived 

generality compared with the generic option. Our reasoning follows from the fact that the Internet allows 

consumers to access detailed and specific information relating to products and services (Hoffman and 

Novak 1996). 

For almost all the products studied, the brand increased all three dimensions of consumers’ 

perceived intangibility. A possible explanation for these unexpected results stems from the Internet’s 

ability to reduce generality and mental intangibility, especially in a generic context, by efficiently 

distributing product-related information to consumers (Hoffman and Novak 1996). All branded goods 

were also rated as more physically intangible than the generic counterparts. Regarding the internet 

browser, Goutaland (1999) found as physically intangible as other services (i.e., charter flight for a 

vacation). It is therefore not surprising to find the Internet browser mimicking the services studied on the 

brand-generic category comparison. Respondents perceived all three goods (i.e., a pair of Jeans, a 

computer, and a CD) as more physically intangible in a branded context as opposed to a generic one. This 

phenomenon may be partly due to the underlying impact of the Internet, which increases physical 

intangibility and prevents consumers from having a sensory experience. Respondents’ online purchase 

experience may also explain this finding for their online purchase frequency of generic products was 

higher than that of branded ones. 

5.2. Intangibility, Knowledge, and Involvement: Difficulty of Evaluation (H4a, H5a, and H6a) 

Finding support for H4a, H5a, and H6a, we conclude that the three dimensions of intangibility, 

knowledge, and involvement are positively related with difficulty of evaluation. The real contribution of 

94

these findings rests on the positive association between all three dimensions of intangibility and difficulty 

of evaluation, something not previously demonstrated. 

5.3. Intangibility, Knowledge, and Involvement: Perceived Risk (H4b, H5b, and H6b) 

Like Goutaland (1999), who found a positive relationship between mental intangibility and overall 

perceived risk, we uncovered links between mental intangibility and all five types of risk (H4b). The 

finding that generality only affects social and psychological risk (H5b) may also explain why it had no 

significant effect on overall perceived risk in Goutaland’s study. Though a positive relationship between 

physical intangibility and risk was expected (e.g., Mitchell and Greatorex 1993), Goutaland (1999) found 

no such association. Similar findings were reproduced here across all four types of risk, except social risk 

(H6b). Mental intangibility, which is more apt to impact perceived risk than generality or physical 

intangibility, should therefore be targeted to reduce perceived risk associated with virtual and traditional 

products/services. 

5.4. Intangibility, Difficulty of Evaluation, and Perceived Risk: Brand and Generic Product Categories 

(H7a and H7b) 

Since the features/characteristics of generic products/services are harder to describe and to 

physically grasp than those of branded ones, the effects of generality and physical intangibility on 

evaluation difficulty and risk were consistently greater in the generic context. The impact of mental 

intangibility on both difficulty of evaluation and perceived risk, was systematically greater at the brand 

level as opposed to the generic one. This finding is contrary to the general belief that brands tangibilize 

goods/services and thus help reduce evaluation difficulty and perceived risk. 

5.5. Knowledge and Involvement (H8 to H13) 

5.5.1. Knowledge and Involvement: Direct Relationships (H9a, H9b, H12a, and H12b) 

Similar to other studies, we found full or partial support for all the hypotheses linking knowledge 

and involvement with difficulty of evaluation and perceived risk (i.e., H9a, H9b, H12a, and H12b) (e.g., 

Goutaland 1999). The relationship between involvement and time risk proved to be an important 

exception. Consumer’s willingness to spend time on purchase-related information search translates into 

95

greater involvement with the purchase and less time risk. Being of greater interest, we discuss the 

moderating effects of knowledge and involvement in more detail in the following sections. 

5.5.2. Moderating Effects of Knowledge and Involvement (H8a, H8b, H11a, and H11b) 

As expected, knowledge moderated the link between mental intangibility and difficulty of 

evaluation (H8a). This finding is in line with many scholars who agree that consumer knowledge is 

accumulated from product-related information search and experience (Anderson et al. 1979; Bettman and 

Park 1980). Associated with information search and product/service exposure, the mental dimension of 

consumer experience serves to make good/service evaluations less difficult. It is possible for physical 

intangibility and generality to not affect the evaluation process since neither increase consumers’ 

processing efforts nor reduce their objective knowledge to perform the evaluation (Wendler 1983). 

Knowledge also moderated the relationships between: physical intangibility and social risk, generality 

and time risk, and mental intangibility and social and psychological risk (H8b). While consumers’ 

objective knowledge may be present, their subjective knowledge (Cox 1967) appears to be affected by the 

intangibility of the good/service under consideration. Consequently, this subjective uncertainty results in 

greater perceived risk of the purchase decision. 

Although we expected involvement to moderate the relationship between intangibility and 

difficulty of evaluation, such was not the case (H11a). This result is contrary to other studies where 

involvement acted as moderator between physical intangibility and difficulty of evaluation (see 

McDougall 1987), and between mental intangibility, generality, and difficulty of evaluation (see 

Goutaland 1999). Since purchase involvement encompasses information search and time spent in making 

a correct decision, consumers probably spend more time and effort in screening information on the 

Internet (Clarke and Belk 1978). As a result, the proposed moderating effects of involvement are much 

less influential. Partially supporting H11b, involvement moderated the links between the various 

dimensions of intangibility and risk. It is said that the more involved a consumer is with a good/service, 

the more perceptive s/he will be regarding attribute differences, product importance, and brand 

96

commitment (Howard and Sheth 1969). Similarly, intangibility appears to influence consumers’ risk 

perceptions with physical intangibility having the most widespread effect. 

5.5.3. Moderating Effects of Knowledge and Involvement: Brand and Generic Product Categories (H10a, 

H10b, H13a, and H13b) 

Knowledge acted as a moderator between mental intangibility and difficulty of evaluation. Since 

this moderating effect was greater at the brand level than the generic one, H10a was not supported. 

Finding no support for our hypothesis (H10a), led us to speculate that respondents’ prior experience with 

the branded and generic products/services chosen was enough to suppress any brand/generic differences, 

especially those relating to difficulty of evaluation. Respondents’ level of involvement with the chosen 

goods/services may have been too weak for the moderating effect to surface in the brand/generic context. 

Thus, H13a was partially supported with involvement only moderating the link between physical 

intangibility and difficulty of evaluation, and this effect being greater in the generic context. 

Contrary to difficulty of evaluation, respondents’ level of perceived risk vis-à-vis the chosen 

goods/services was ubiquitous. This led to us finding partial support for both H10b and H13b with 

knowledge and involvement acting as moderators between various dimensions of intangibility and risk, 

and the effects being greater in the generic context. These results are to be expected for brands are known 

to reduce perceived risk (Roselius 1971; Sheth and Venkatesan 1968; Wernerfelt 1988). 

6. Theoretical and Managerial Implications 

Since the tested model shows that the three dimensions of intangibility all affect consumer 

difficulty of evaluation and perceived risk, marketers should rate their goods and services on an 

intangibility scale and/or position their goods/services on an intangibility map, determine the most 

influential dimension of intangibility, and ultimately design marketing strategies that will lessen 

perceived risk and simplify evaluations. In our study, for example, the Internet browser was perceived as 

more physically, as opposed to mentally, intangible suggesting that marketers could lessen consumers’ 

perceived risk and evaluation difficulty by decreasing the perceived physical intangibility of their 

offerings. For other goods/services, marketers should also bear in mind that mental intangibility, along 

97

with generality, may also affect consumers’ perceptions of risk and evaluation difficulty. Hence, their 

corresponding marketing strategies should decrease, or at least, not increase mental intangibility and 

generality. 

An important theoretical contribution of this study is that brands function to reduce intangibility in 

both offline and online purchase situations. While the branding strategies employed by service companies 

are still useful when moving into the world of e-commerce, such is not the case for products purchased 

online. In addition, the five types of perceived risk turn out to have different relationships with the 

different variables studied here. For example, involvement interacts with mental intangibility to reduce 

social and psychological risk, while knowledge interacts with generality to increase perceived time risk. 

By mapping their goods/services according to the different types of risk associated with each, marketers 

will be better able to communicate their risk-reduction strategies. 

The appearance of very intangible goods and very tangible services brings forth challenges for all 

marketers as to how best to communicate the perceived intangible attributes of a virtual good/service, 

how to position the product/service, and how to reduce consumers’ purchase-related perceived risks and 

evaluation difficulty. These challenges suggest that branded and generic goods/services, and their 

intangibility-related consequences, should be studied in greater detail. 

7. Limitations and Directions for Future Research 

Even though students are generally familiar with online purchases, a more diverse sample with 

online purchase experience should be tapped. More variety in the goods/services used would also help 

matters especially since the intangibility exhibited by some virtual products (e.g., Internet browser) is 

similar to that displayed by services. Generally, the distribution of online purchase frequency of branded 

products was considerably lower than that of generic products, which may explain why the means of 

branded products were higher than those of generic products. Future studies should thus control for these 

differences in distribution. 

This study is meaningful in that it confirms the three-dimensional intangibility construct and finds 

brands to be major intangibility-reducers especially for services. While brands reduce consumers’ 

98

perceived physical intangibility, mental intangibility, and generality for services such is not the case for 

goods. Future research should consider how the Internet will affect consumers’ perceived intangibility of 

‘virtual’ goods/services be they branded or not. Brands also reduce consumers’ perceived evaluation 

difficulty and risk, especially for goods/services with high perceived physical intangibility and generality. 

Since mental intangibility associated with branded goods/services only helps reduce perceived social risk, 

future research should explore whether other factors influence mental intangibility. While brands reduce 

the moderating effects of knowledge and involvement, this finding does not hold true in all cases. Future 

research should thus explore the reasons behind these results by studying a varied array of goods/services. 

Knowledge and involvement also directly affect consumer evaluation difficulty and perceived risk and act 

as moderators. Future research should consider the moderating effects of knowledge and involvement 

while controlling for online offerings (i.e., online goods/services held as covariates). 

References 

Alba JW, Hutchinson JW. Dimensions of consumer expertise. Journal of Consumer Research 

1987;13(March):411-454. 

Anderson RD, Engledow JL, Becker H. Evaluating the relationships among attitude toward business 

product satisfaction, experience, and search effort. Journal of Marketing Research 

1979;16(August):394-400. 

Barach JA. Advertising effectiveness and risk in the consumer decision process. Journal of Marketing 

Research 1969;August:314-20. 

Bateson JEG. Why we need service marketing. Conceptual and Theoretical Developments in Marketing, 

Proceedings Series of the American Marketing Association. Chicago, IL: American Marketing 

Association, 1979. p. 131-146. 

Bauer RA. Consumer behavior at risk taking. In: Cox DF, editor. Risk Taking and Information Handling 

in Consumer Behavior. Cambridge, MA: Harvard University Press. 1960. p. 23-33. 

Bauer RA. Risk handling in drug adoption: the role of company preference. Public Opinion Quarterly 

1961(Winter):546-559. 

Bebko CP. Service intangibility and its impact on consumer expectations of service quality. 

Journal of Services Marketing 2000;14(1):9-26. 

Berry LL. Services marketing is different. Business 1980;(May-June):16-23. 

Berry LL. Cultivating service brand equity. Journal of the Academy of Marketing Science 

2000;28(1):128-37. 

99

Berry LL, Clark T. Four ways to make services more tangible. Business 1986;36(4):53-54. 

Bettman JR. Consumer psychology. In: Lutz R, editor. Advances in Consumer Research, vol. 13. 

Provo, UT: Association for Consumer Research, 1986. p. 257-289. 

Bettman JR, Park CW. Effects of prior knowledge and experience and phases of the choice 

process on consumer decision process: A protocol analysis. Journal of Consumer Research 

1980;7(3):234-248. 

Biehal GJ. Consumers’ prior experiences and perceptions in auto repair choice. Journal of 

Marketing 1983;47(Summer):82-91. 

Bloch PH. An exploration into scaling of consumers’ involvement with a product class. In: 

Monroe KB, editor. Advances in Consumer Research, vol. 8. Provo, UT: Association for 

Consumer Research, 1981. p. 61-65. 

Breivik E, Troye SV, Olsson UH. Dimensions of intangibility and their impact on product 

evaluation. Working Paper Poster Session. Montreal, Canada: Association for Consumer 

Research, 1998. 

Brucks M. The effects of product knowledge on information search behavior. Journal of 

Consumer Research 1985;12(June):1-16. 

Brucks M. A Typology of Consumer Knowledge Content. In: Lutz R, editor. Advances in 

Consumer Research, vol. 13. Provo, UT: Association for Consumer Research, 1986. p. 58- 

63. 

Brucks M, Mitchell AA. Knowledge structures, production systems and decision structures. In: 

Monroe KB, editor. Advances in Consumer Research, vol. 8. Provo, UT: Association for 


Clarke K, Belk RW. The effect of product involvement and task definition on anticipated 

consumer effort. In: Hunt K, editor. Advances in Consumer Research, vol. 5. Provo, UT: 

Association for Consumer Research, 1978. p. 303-18. 

Cox DF. Risk Taking and Information Handling in Consumer Behavior. Boston, MA: Harvard 

University Press, 1967. 

Cox DF, Rich SV. Perceived risk and consumer decision making – the case of telephone 

shopping. In: Cox DF, editor. Risk Taking and Information Handling in Consumer Behavior. 

Cambridge, MA: Harvard University Press. 1967. p. 487-506. 

Cunningham SM. The major dimensions of perceived risk. In: Cox DF, editor. Risk Taking and 

Information Handling in Consumer Behavior. Cambridge, MA: Harvard University Press. 

1967. p. 23-33. 

Dubé-Rioux L, Regan DT, Schmitt BH. The cognitive representation of services varying in 

concreteness and specificity. In: Goldberg M, Gorn G, Pollay R, editors. Advances in 

100


865. 

Edgett S, Parkinson S. Marketing for service industries – a review. The Service Industries 

Journal 1993;13(9):19-36. 

Engel F, Blackwell RD. Consumer Behavior, 4 th ed. New York, NY: The Dryden Press.1982. 

Erdem T. 1998. An empirical analysis of umbrella branding. Journal of Marketing Research 

1998;35(3):339-51. 

Goutaland C. Product and service intangibility: a study of its dimensions and consequences on 

product/service evaluation. Unpublished Master’s Thesis, Concordia University, Montreal, 

Canada. 1999. 

Havlena WJ, DeSarbo WS. On the measurement of perceived consumer risk. Decision Sciences 

1990;22:927-939. 

Hirschman EC. Attributes of attributes and layers of meaning. In: Monroe KB, editor. Advances 

in Consumer Research, vol. 8. Provo, UT: Association for Consumer Research, 1981. p. 7- 

12. 

Hoffman DL, Novak TP. Marketing in hypermedia computer-mediated environments: 

conceptual foundations. Journal of Marketing 1996;60(3):50-53. 

Howard JA. Consumer Behavior: Application of Theory. New York, NY: John Wiley. 1977. 

Howard JA, Sheth JN. The Theory of Buyer Behavior. New York, NY: John Wiley. 1969. 

Jacoby J, Kaplan L. The components of perceived risk. In: Venkatesan M, editor. Advances in 

Consumer Research, Proceedings of the 1972 Conference. Provo, UT: Association for 


Johnson EJ, Fornell C. The nature and methodological implications of the cognitive 

representation of products. Journal of Consumer Research 1987;14(September):214-228. 

Johnson EJ, Russo EJ. Product familiarity and learning new information. Journal of Consumer 

Research 1984;11(June):542-550. 

Kaplan L, Szybillo G, Jacoby J. Components of perceived risk in product purchase: a crossvalidation. 

Journal of Applied Psychology 1974;59:287-291. 

Kotler P, Bowen J, Makens J. Marketing for Hospitality and Tourism. Upper Saddle River, NJ: 

Prentice-Hall. 1999. 

Krugman HE. Memory without recall, exposure without perception. Journal of Advertising 

Research 1977;17(August):7-12. 

101

Laroche M, Bergeron J, Goutaland C. A three-dimensional scale of intangibility. Journal of 

Service Research 2001;4(1):26-38. 

Laurent G, Kapferer J-N. Measuring consumer involvement profiles. Journal of Marketing 

Research 1985;22(February):41-53. 

Lovelock C. Services Marketing: People, Technology, Strategy. Fourth edition, Upper Saddle 

River, NJ: Prentice-Hall. 2001. 

McDougall GH. Determinants of ease of evaluation: products and services compared. Canadian 

Journal of Administrative Sciences 1987;4(4):426-446. 

McDougall GH, Snetsinger DW. The intangibility of services: measurement and competitive 

perspectives. Journal of Services Marketing 1990;4(4):27-40. 

Mitchell VW, Greatorex M. Risk perception and reduction in the purchase of consumer services. 

The Service Industries Journal 1993;13(4):179-200. 

Mittal BL. A theoretical analysis of two recent measures of involvement. In: Wilkie WL, editor. 

Advances in Consumer Research, vol. 16. Provo, UT: Association for Consumer Research, 

1989. p. 697-702. 

Mittal BL, Baker J. Advertising strategies for hospitality services. Cornell Hotel and Restaurant 

Administration Quarterly 2002;43(2):51-63. 

Murray KB. A test of services marketing theory: consumer information acquisition activities. 

Journal of Marketing, 1991;55:10-25. 

Murray KB, Schlacter JL. The impact of services versus goods on consumer’s assessment of 

perceived risk and variability. Journal of the Academy of Marketing Science 1990;18(1):51- 

65. 

Oliver RL, Bearden WO. The role of involvement in satisfaction processes. Advances in 


255. 

Park CW, Mothersbaugh DL, Feick L. Consumer knowledge assessment. Journal of Consumer 

Research 1994:21(June):71-82. 

Perry M, Hamm BC. Canonical analysis of relations between socioeconomic risk and personal 

influence in purchase decisions. Journal of Marketing Research 1968;6(August):351-354. 

Rathmell JM. Marketing in the Service Sector. Cambridge, MA: Winthrop Publishers. 1974. 

Reddy AC, Buskirk BD, Kaicker A. Tangibilizing the intangibles: some strategies for service 

marketing. Journal of Services Marketing 1993;7(3):13-17. 

102

Roselius TR. Consumer rankings of risk reduction methods. Journal of Marketing 

1971;31(1):56-61. 

Rust RT, Zahorik AJ, Keiningham TL. Readings in Service Marketing. New York, NY: Harper 

Collins College Publishers. 1996. 

Sheth J, Venkatesan M. Risk reduction processes in repetitive consumer behavior. Journal of 

Marketing Research 1968;(August):307-10. 

Shostack GL. Breaking free from product marketing. Journal of Marketing 1977;(April):73-80. 

Stone RN, Grønhaug K. Perceived risk: further considerations for the marketing discipline. 

European Journal of Marketing 1993;27(3):39-50. 

Surprenant CF, Solomon MR. Predictability and personalization in the service encounter. Journal 

of Marketing 1987;51(April):86-96. 

Van Dierdonck R. Success strategies in a service economy. European Management Journal 

1992;10(3):365-364. 

Wendler ER. Consumer Information and Confidence: Moderating Effects of Perceived 

Comprehension and Risk. In: Bagozzi R, Tybout A, editors. Advances in Consumer 

Research, vol. 10. Provo, UT: Association for Consumer Research, 1983. p. 364-369. 

Wernefelt B. Umbrella branding as a signal of new product quality: an example of signaling by 

posting a bond. Rand Journal of Economics 1988;19(4):458-66. 

Zaichkowsky JL. Measuring the involvement construct. Journal of Consumer Research 

1985;12(December):341-352. 

Zeithaml VA. How consumer evaluation processes differ between goods and services. In: 

Donnelly JH Jr, George WR, editors. Marketing of Services, AMA’s Special Conference on 

Services Marketing, American Marketing Association, 1981. p. 186-190. 

Zeithaml VA, Berry LL, Parasuraman A. Problems and strategies in services marketing. Journal 

of Marketing 1985;49(Spring):33-46. 

Zeithaml VA, Bitner MJ. Services Marketing: Integrating Customer Focus Across Firms, New 

York, NY: McGraw-Hill. 2000. 

103

Questionnair 

es 

Questionnaire 

1 

Questionnaire 

2 

Questionnaire 

3 

Questionnaire 

4 

Questionnaire 

5 

Questionnaire 

6 

Questionnaire 

7 

Questionnaire 

8 

Questionnaire 

9 

Questionnaire 

10 

Questionnaire 

11 

Questionnaire 

12 

Questionnaire 

13 

Questionnaire 

14 

Questionnaire 

15 

Questionnaire 

16 

Table 1 

The sixteen versions of the questionnaire 

Products and service mix included in the 

questionnaire 

Online or 

offline 

purchase 

Brand or 

generic 

category 

Levi’s jeans, Pizza Hut’s pizza, Netscape software Online Brand 

IBM computer, Royal Bank checking account, 

Beatles’ CD 

Online Brand 

Pair of jeans, Pizzeria dinner, Internet browser Online Generic 

Computer, Checking account, CD Online Generic 

Levi’s jeans, Pizza Hut’s pizza, Netscape software Offline Brand 

IBM computer, Royal Bank checking account, 

Beatles’ CD 

Offline Brand 

Pair of jeans, Pizzeria dinner, Internet browser Offline Generic 

Computer, Checking account, CD Offline Generic 

Netscape software, Pizza Hut’s pizza, Levi’s jeans Online Brand 

Beatles’ CD, Royal Bank checking account, IBM 

Computer 

Online Brand 

Internet browser, Pizzeria dinner, Pair of jeans Online Generic 

CD, Checking account, Computer Online Generic 

Netscape software, Pizza Hut’s pizza, Levi’s jeans Offline Brand 

Beatles’ CD, Royal Bank checking account, IBM 

computer 

Offline Brand 

Internet browser, Pizzeria dinner, Pair of jeans Offline Generic 

CD, Checking account, Computer Offline Generic 

Table 2 

Number of respondents according to online/offline and brand/generic categories 

Generic Brand Total 

104

Online 

Offline 

Total 

196 195 391 

195 197 392 

391 392 783 

105

Table 3 

Results of factor analysis 

Construct Items % of 

Varianc 

Intangibilit 

y 

Perceived 

risk 

Difficulty of 

evaluation 

Knowledge 

Involvemen 

t 

Physical intangibility (factor 1) 

PHYS11: I can physically grasp ____ 

PHYS12: _____ are physically very tangible 

PHYS10: ____ is very easy to see and touch 

Mental intangibility (factor 2) 

MENTAL20: This is not the sort of ___ that is easy to picture 

MENTAL19: This is a difficult ____ to think about 

MENTAL18: I need more information about ______ to get a clear idea (image) of what it is 

Generality (factor 3) 

GEN14: It is not difficult to give a precise description of a ____ 

GEN15: It is easy to describe many features related to a ____ 

GEN13: I could easily explain many features associated with ____ 

Time risk (factor 4) 

TRISK38: Purchasing a ____ will involve important time losses 

TRISK37: Purchasing a ___ will lead to an inefficient use of my time 

TRISK39: The demands on my schedule are such that purchasing a ___ concerns me because it 

would impose even greater time pressure on me. 

Social risk (factor 5) 

SRISK44: If I used a ____, I would be held in higher esteem by my family 

SRISK43: If I used a ____, I would be held in higher esteem by my friends 

SRISK45: Purchasing a ____ within the next twelve months would cause me to be considered as 

foolish by some people whose opinion I value 

Performance risk (factor 6) 

PRISK41: As I consider the purchase of a ___ in the near future, I worry about whether it will 

really “perform” as well as it is supposed to 

PRISK42: The thought of purchasing a _____ causes me to be concerned for how really reliable 

that product will be 

PRISK40: If I were to purchase a ____ within the next twelve months, I would be concerned that 

the brand will not provide the level of benefits that I would be expecting 

Psychological risk (factor 7) 

YRISK48: The thought of purchasing a ____ causes me to experience unnecessary tension 

YRISK47: The thought of purchasing a ____ makes me feel psychologically uncomfortable 

YRISK46: The thought of purchasing a ____ gives me a feeling of unwanted anxiety 

Financial risk (factor 8) 

FRISK35: Purchasing a ___ could involve important financial losses 

FRISK34: If I bought a ___ for myself within the next twelve months, I would be concerned that 

this financial investment would be wise 

FRISK36: If I bought a ___ for myself within the next twelve months, I would be concerned that 

I would not get my money’s worth 

Difficulty of evaluation (factor 9) 

Given that I have to buy a ___ , evaluating a ___ will be: 

DIFF24: Very complicated to not complicated at all 

DIFF23: Very complex to very simple 

DIFF22: Very problematic to not problematic at all 

DIFF21: Very difficult to very easy 

Knowledge (factor 10) 

KNOW6: Compared to experts in this area, my knowledge of a ______ is 

KNOW4: Would you consider yourself uninformed or informed about a____ 

KNOW3: In general, my knowledge of a ____ is 

KNOW5: Compared to my friends and acquaintance, my knowledge of a ____ is 

INFO8: The global information search I have performed on ______ is 

EXP9: I do not have much experience purchasing _____ 

EXP7: I use ___ (never to very often) 

Involvement (factor 11) 

INV26: I perceive a ____ is: Very significant to very insignificant 

INV25: I perceive a ____ is: Very important to very unimportant 

INV27: I perceive a ____ is: Very valuable to not valuable at all 

INV28: A ____ is: Matters a lot to me to does not matter to me 

106 

e 

5.64 

4.13 

2.60 

25.46 

6.79 

2.42 

2.31 

1.91 

5.10 

16.40 

9.11 

Loadin 

g 

.941 

.898 

.797 

.879 

.873 

.663 

.925 

.907 

.850 

.970 

.933 

.775 

.968 

.960 

.548 

.939 

.814 

.596 

.919 

.914 

.913 

.900 

.863 

.758 

.942 

.936 

.902 

.861 

.842 

.835 

.807 

.800 

.719 

.583 

.536 

.951 

.941 

.923 

.838 

Alph 

a 

.83 

.91 

.87 

.94 

.87 

.91 

.97 

.91 

.95 

.89 

.94

INV29: I perceive a ____ is: Means a lot to me to means nothing to me .828 

Extraction Method: Principal Component Analysis; Rotation: Oblimin with Kaiser Normalization. 

107

Table 4 

Summary of results 

108

Hypotheses Results 

H1: The degree of generality associated with generic category- H1 is supported for services 

level choices is higher than it is for brand-level choices. H1 is not supported for goods 

H2: The degree of mental intangibility associated with H2 is supported for services 

generic category-level choices is higher than it is for brandlevel 

choices. 

H2 is not supported for goods 

H3: The degree of physical intangibility associated with 

generic category-level choices is a) higher than it is for 

brand-level choices of services and b) similar to it for 

brand-level choices of products. 

H3a is supported for services 

H3b is not supported for goods 

H4: The more mentally intangible a product/service is, a) the more H4a is supported 

difficult it is to evaluate and b) the more risky it is perceived. H4b is supported (all 5 types of risk) � 

H5: The more general a product/service is, a) the more difficult it is H5a is supported 

to evaluate and b) the more risky it is perceived. H5b is partially supported (social and psychological risk) � 

H6: The more physically intangible a product/service is, a) the H6a is supported 

more difficult it is to evaluate and b) the more risky it is perceived. H6b is not supported 

H7: The impact of intangibility on a) difficulty of evaluation and b) H7a is Physical intangibility � 

perceived risk is greater for generic-category level choice than partially Generality � 

brand-level choice. 

supported Mental intangibility � 

H7b is Physical intangibility 

partially (social, financial, and performance risk) � 

supported Generality (all 5 types of risk) � 

Mental intangibility (social risk) � 

H8: Knowledge moderates the relationships between the three H8a is Physical intangibility � 

dimensions of intangibility (i.e., generality, physical intangibility partially Generality � 

and mentally intangibility) and a) difficulty of evaluation and b) 

perceived risk. 

supported 

H8b is 

Mental intangibility � 

Physical intangibility (social risk) � 

partially Generality (time risk) � 

supported Mental intangibility (social and psychological risk) � 

H9: The more knowledgeable a consumer perceives himself to be H9a is supported 

regarding a product/service a) the less difficult it is to evaluate it 

and b) the less risky it is perceived. 

H9b is partially supported 

(psychological, financial, performance, and time risk) � 

H10: The moderating effects of knowledge on a) difficulty of H10a is not supported 

evaluation and b) perceived risk at the generic level are greater than 

those at the brand level. 

H10b is 

partially 

supported 

Physical intangibility (social and performance risk) � 

Generality (social and psychological risk) � 

Mental intangibility (performance risk) � 

H11: Involvement moderates the relationships between the three 

dimensions of intangibility (i.e., generality, physical intangibility, 

and mental intangibility) and a) difficulty of evaluation and b) 

perceived risk. 

H12: The more involving a product/service is for a consumer, a) 

the more difficult it is to evaluate it and b) the more risky it is 

perceived. 

H13: The moderating effects of involvement on a) difficulty of 

evaluation and b) perceived risk at the generic level are greater than 

those at the brand level. 

�: hypothesis supported �: hypothesis not supported 

H 11a is not supported 

H11b is 

partially 

supported 

Physical intangibility (social, psychological, financial, 

and performance risk) � 

Generality (social and time risk) � 

Mental intangibility (social and psychological risk) � 

H12a is supported 

H12b is partially supported (social and financial risk) � 

H13a is Physical intangibility � 

partially Generality � 

supported Mental intangibility � 

H13b is Physical intangibility 

partially (psychological, financial, and time risk ) � 

supported Generality � 

Mental intangibility 

(social, psychological, financial, and time risk) � 

109

Table 5 

Stepwise multiple regressions (H4a, H5a, H6a, H9a, and H12a) 

Intangibility, knowledge, and involvement with difficulty of evaluation 

Standardized 

Coefficients 

T-Values Level of 

Significance 

(Constant) 17.862 .000 

Physical 

intangibility 

.051 2.157 .015 

Mental 

intangibility 

.079 3.600 .000 

Generality .178 6.551 .000 

Knowledge -.207 -8.429 .000 

Involvement .094 4.426 .000 

Table 6 

Stepwise multiple regressions (H4b, H5b, H6b, H9b, and H12b) 

Intangibility, knowledge, and involvement with perceived risk 

Time risk 

Social risk 

Psychological risk 

Adjusted R 2 = .147 

F = 81.85 

Financial risk 

Performance risk 

Adjusted R 2 = .066 Adjusted R 

F = 56.74 

2 = .032 Adjusted R 

F = 20.31 


F = 29.58 


F = 41.23 

2 = .034 

F = 42.05 

Physical intangibility -.092* 

Mental intangibility .118 * .160* .137* .134* .120* 

Generality .067** .052*** 

Knowledge -.142* -.045*** -.161* -.110* 

Involvement -.094* .099* .090* 

Level of significance: * p ≤.001, ** p≤.01, *** p≤.05 

Table 7 

Stepwise multiple regressions (H8a and H11a) 

Moderating effects of knowledge and involvement: intangibility and difficulty of evaluation 

110

(Constant) 

Knowledge 

Knowledge and 

Mental 

intangibility 

Standardized 

Coefficients 

T-Values Level of 

Significance 

13.014 .000 

-.140 -3.811 .000 

-.112 -2.450 .007 

111 


F = 69.36

Table 8 

Stepwise multiple regressions (H8b and H11b) 

Moderating effects of knowledge and involvement: intangibility and perceived risk 

Time risk 

Social risk 

Psychological risk 

Financial risk 

Performance risk 

Adjusted R 2 = .077 Adjusted R 

F = 33.88 


F = 16.42 


F = 14.73 


F = 30.24 

2 = .041 

F = 21.32 

Knowledge -.204* -.194* -.134* -.160* -.102* 

Knowledge and 

Physical intangibility 

.296** 

Knowledge and 


.296* .130** 

Knowledge and 

Generality 

.086*** 

Involvement .089** .247* .198* .210* .110* 

Involvement and 


-.193* -.208* -.263* -.253* 



-.222* -.135** 


Generality 

-.292* .097* 

Level of significance: * p≤.001, ** p≤.01, *** p≤.05 

Table 9 

Stepwise multiple regressions (H7a) 

Intangibility and difficulty of evaluation: brand and generic product categories 

Generic 


F = 82.41 

Physical intangibility .064*** 

Difficulty of evaluation 

Brand 


F = 97.57 

Mental intangibility .168* 

Generality .311* .267* 


112

Physical 

intangibility 

Mental 

intangibility 

Table 10 

Stepwise multiple regressions (H7b) 

Intangibility and perceived risk: brand and generic product categories 

Generic 

Adjusted 

R 2 = .036 

F = 22.86 

Time risk Social risk Psychological risk Financial risk Performance risk 

Brand 

Adjusted 

R 2 = .035 

F = 22.37 

Generic 

Adjusted 

R 2 = .04 

F = 17.12 

- 

.107** 

Brand 

Adjusted 

R 2 = .013 

F = 8.81 

- 

.064*** 

Generic 

Adjusted 

R 2 = .043 

F = 27.59 

Brand 

Adjusted 

R 2 = .029 

F = 36.04 

Generic 

Adjusted 

R 2 = .039 

F = 16.76 

Brand 

Adjusted 

R 2 = .037 

F = 45.56 

Generic 

Adjusted 

R 2 = .025 

F = 11.06 

Brand 

Adjusted 

R 2 = .041 

F = 26.12 

-.089** -.091** . 060*** 

.124* .135* .161* .130* .127* .173* .126* .193* .112* .176* 

Generality .104* . 086** .127* .123* .159* .126* 


Table 11 

Stepwise multiple regressions (H10a and H13a) 

Moderating effects of knowledge and involvement on difficulty of evaluation: 

brand and generic product categories 

Generic 


F = 56.21 

Knowledge -.311* 

Knowledge and 


Knowledge and 


Knowledge and 

Generality 

Involvement .181* 


-.207** 


Difficulty of evaluation 

Brand 


F = 70.65 

-.144* 




Generality 


113

114

Table 12 

Stepwise multiple regressions (H10b and H13b) 

Moderating effects of knowledge and involvement on perceived risk: 

Generic 

Adjusted 

R 2 = .086 

F = 19.35 

brand and generic product categories 

Time risk Social risk Psychological risk Financial risk Performance risk 

Brand 

Adjusted 

R 2 = .074 

F = 32.26 

Generic 

Adjusted 

R 2 = .064 

F = 12.46 

Brand 

Adjusted 

R 2 = .049 

F = 13.01 

Generic 

Adjusted 

R 2 = .069 

F = 13.43 

Brand 

Adjusted 

R 2 = .029 

F = 36.04 

Generic 

Adjusted 

R 2 = .097 

F = 21.87 

Brand 

Adjusted 

R 2 = .043 

F = 27.21 

Generic 

Adjusted 

R 2 = .061 

F = 19.99 

Brand 

Adjusted 

R 2 = .063 

F = 27.40 

Knowledge -.153* -.200* -.152** -.195* -.136* -.243* -.088** -.113*** 

Knowledge and 


-.113* -.066*** 

Knowledge and 


.130* .284* .301* 

- 

.141**** 

Knowledge and 

Generality 

.136* .064*** 

Involvement .187* -.147* .246* .276* .250* .335* 



-.415* -.101** -.344* -.316* -.238* 




Generality 

-.156*** -.371* -.133** -.187** 

- 

.139**** 

Level of significance: * p≤.001, ** p≤.01, *** p≤.05, **** p≤.10 

115

Intangibility 



Generality 

Legend: 

Positive effects (H4, H5, H6, and H12) 

Figure 1 

Proposed model 

Knowledge 

Involvement 

Moderating effects of knowledge and involvement (H8 and H11) 

Negative effects (H9) 

Difficulty of 

evaluation 

Perceived risk 

Shaded areas represent differences between brand and generic perspectives (H1, H2, H3, H7, H10, and 

H13) 

116

Exploring the WOW in Online Auction Feedback 

Bruce D. Weinberg, Bentley University 

Lenita Davis , The University of Alabama 

ABSTRACT 

Ebay, the leading online-retailer, has captured the fascination of both American consumers and 

corporations, as it has garnered approximately 45% of all gross online retail sales, $5.3 billion 

out of $11.9 billion, during the first quarter of 2003. George Day has described eBay “as perhaps 

the most successful of all the breakthrough applications on the Internet.” While there are many 

reasons cited for the success of online auctions, the reputation system, or Feedback Forum as it is 

called on eBay, has been considered the most critical element. Some auction studies have 

investigated the impact of seller-reputation on consumer/bidder behavior and auction results; 

however, they have considered only aggregate or summary measures of reputation, such as the 

overall feedback rating. This research extends our understanding of online-auction reputation 

systems by a) classifying online-auction feedback as a type of “word-of-web” (WOW), 

specifically, rating-and-review word-of-web, b) identifying a property that is unique to online 

auctions -- “bidirectionality,” and c) discovering that, in addition to the overall feedback rating, 

consumers/bidders process several feedback profile elements/details, such as individual-auction 

feedback reviews and comments, when evaluating sellers and making bidding decisions. 

INTRODUCTION 

Ebay, the leading online-retailer, has captured the fascination of both American consumers and 

corporations (Adler 2002). With 68.8 million members, it has garnered approximately 45% of all 

gross online retail sales, $5.3 billion out of $11.9 billion, during the first quarter of 2003 (eBay 

2003, US Census Bureau 2003). And the future of Internet auctions looks bright as eBay 

financial results for Q1 2003 compared to Q1 2002 reported increases in net revenues and GAAP 

117

net income of 94% and 119%, respectively; also sales and marketing expenses as a percentage of 

net revenues decreased to 26% from 30% during that same time period (eBay 2003). 

Prominent marketers have proclaimed the importance of studying online auctions; and 

practitioners have pointed to eBay in particular as the torch bearer for success in e-commerce. 

For example, Chakravarti et al (2002) strongly recommends that marketers take a stronger 

interest in online auctions by emphasizing “despite the growing interest in traditional and 

Internet auctions, the marketing literature is sparse;” and, George Day has described eBay “as 

perhaps the most successful of all the breakthrough applications on the Internet” (CNET 2002). 

While there are many reasons cited for the success of online auctions, the reputation system, or 

Feedback Forum as it is called on eBay, has been considered the most critical element (Grant 

2002, Hof 2001). EBay’s Feedback Forum is where buyers and sellers can rate and read about 

each others’ past transaction experiences. Each eBay member has a Feedback Profile which 

tracks and maintains several pieces of reputation-related information about the member, 

including individual-auction feedback reviews, which consist of a satisfaction/dissatisfaction 

rating and comments by others who have participated in an online-auction exchange with this 

member. It also has an overall feedback rating which is, in essence, an aggregate reputation score 

(for an example, see Figure 1). 

FIGURE 1 about here 

Most online auction studies have focused on eBay or used eBay-generated data; and, those that 

have investigated the impact of seller-reputation on consumer/bidder behavior and auction 

results have considered only aggregate or summary measures of reputation, such as the feedback 

rating. Indeed, the overall reputation (score) is an important factor in decision making and 

bidding behavior (e.g., Resnick et al. 2000). However, virtually no research has “gone beneath 

the surface” to investigate consumer utilization of many other Feedback Profile elements/details, 

118

such as the individual-auction feedback reviews, and their impact on consumer decision making 

in online auctions. We believe that such deeper exploration would enhance our understanding of 

eBay’s key-success-factor in specific and online-auction reputation systems in general, and, 

consequently, would reveal insights that could be applied in other online, and perhaps offline, 

retailing applications. 

We address this void in the auction and marketing literature by classifying online-auction 

feedback as a type of “word-of-web,” specifically, rating-and-review word-of-web, and by 

conducting two studies to learn more about consumers’ perceptions and use of online-auction 

feedback. In study one, we analyzed more than 80 websites that reported rating-and-review 

(RR) word-of-web (WOW), developed a general framework that outlines the various 

characteristics of RR WOW, and applied this framework to identify a property that is unique to 

online auctions --“bidirectionality.” In study two, we used eBay to study consumer processing of 

bidirectional online-auction reputation information; we discovered that consumers/bidders 

process several Feedback Profile elements/details beyond the aggregate reputation score of the 

feedback rating when evaluating sellers and making decisions about bidding in their auctions. 

We begin the paper with a brief discussion of the retail online auction market. Next, we review 

relevant literature, including elaboration on the concept of word-of-web communication. Then, 

we describe and discuss the results of study one and study two. Finally, we close with a 

discussion about opportunities for future research. 

“RETAIL” ONLINE AUCTIONS OVERVIEW 

The online auction market is served by several prominent players, including, most notably, 

Amazon, eBay and Yahoo!. eBay, however, is the most dominant player, with more than 

80% market share. Amazon Auctions and Yahoo! Auctions have about 2% share each. Given 

119

this, it is not surprising that eBay is the most referenced online auction player and to many is 

synonymous with the term online auction. 

Many leading manufacturers, retailers, service providers and charitable organizations have 

engaged online auctions: for example, IBM, Dell, Apple, Kodak, Motorola, Boeing, 

Whirlpool, Disney, Sears, The Home Depot, The Sharper Image, ESPN, The American 

Broadcasting Company (ABC), FTD, The American Red Cross and The Make-A-Wish 

Foundation. They use online auctions for a variety of marketing and strategic applications, 

such as estimating demand curves and setting prices, accelerating new product adoption, 

selling refurbished goods and excess inventory, enhancing brand value, and serving new 

segments (for extensive details see Kambil and van Heck 2002). 

Several factors are attributed to the success of eBay’s business model, including: its 

extensive outsourcing and low capital expenditures, the fun factor of participating in the 

auction process, the ability to easily find just about any current or out-of-production product, 

its simple interface design with fast loading webpages, secure technology, a website that 

rarely goes down, and, what has captured the most interest, its having built and maintained a 

strong community. The centerpiece of eBay’s community is its highly regarded Feedback 

Forum, where, according to Gartner’s Walter Janowski, “buyers and sellers are easily able to 

communicate with each other and share ratings and opinions” (Grant 2002). 

Indeed, the general importance of having a strong (overall) feedback rating is widely held. 

However, at the present time, little is known about consumer processing and utilization of the 

various feedback profile elements. But, eBay and other online-auction industry players 

appear to recognize the importance of learning more in this regard, as firms, such as Andale, 

have created tools for managing and tracking feedback, among other auction-related research 

activities. 

120

LITERATURE REVIEW 

Word-of-Web 

Word-of-mouth (WOM) is an extremely important and influential source of information to 

consumers engaged in the buying decision process (e.g., see Gilly et al. 1998, Maxham and 

Netemeyer 2002, Richins 1983, Srinivasan, Anderson, and Ponnavolu 2002). The World 

Wide Web is another effective medium through which consumers spread consumption 

related information among each other. Marketers have referred to consumer information 

diffused to other consumers through this means as electronic or Internet word of mouth 

(Hanson 2000). The spirit of these expressions is fine at face value; however, technically, 

such information transmitted via the Internet is not word-of-mouth. Word-of-mouth is the 

exchange of oral or spoken messages between a source and a receiver concerning the 

purchase of a good or service (Ong 1982) in real time and space (Stern 1994). Word-of-web 

(WOW) is, arguably, a more precise term for describing behavior where product or purchase 

related information is communicated from consumer to consumer via the Web, either at real 

time or not. 

Word-of-web is spread through a variety of electronic forms, such as email, including viral 

marketing (Godin 2001) and e-newsletters (Katz 2002, Scientific American 2001), instant 

messaging, including chat rooms (Gelb and Sundaram 2002), online community or 

discussion forums (Bickart and Schindler 2001, Hagel and Armstrong 1997, Kozinets 2002, 

Rheingold 2000), websites, including weblogs (Cristol 2002, Levy 2002), reviews, and rating 

and reviews (Kuehl 1999). Review information differs from rating-and-review WOW in that 

the latter includes a quantitative measure of some scale type, be it ordinal, interval or ratio, in 

its evaluation of a target object, such as a product. The rating-and-review (RR) form, 

121

frequently referred to as feedback in online auction environments, is available in a number of 

leading online auction sites, and is influential (e.g., Lucking-Reiley 2000). Online auctions 

Though research on auction theory and practice is well-established (e.g., see McAfee and 

McMillan 1987, Milgrom 1989, Milgrom and Weber 1982), very little research about online 

auctions has been published in scholarly journals. This is not necessarily surprising given 

that the first Web-based commercial auction sites were launched in 1995. 

Lucking-Reiley’s (2000) introduction to Internet auctions is significant because he 

investigated 142 Internet auction sites, and provides a comprehensive overview of the 

Internet auction industry, including its early history, business models, goods sold, auction 

formats and options, and concerns about fraud. Herschlag and Zwick (2000) provide a 

review of nonacademic articles which touch on many of the same themes. 

Online auction studies have focused primarily on reputation issues (i.e., online auction 

feedback -- discussed in the next section of this paper), and on auction-listing issues, such as 

the value of providing photographs/images of an auction item (Ottaway, Bruneau and Evans 

2003), setting an opening bid amount (Gilkeson and Reynolds 2003) or setting the bid 

increment amount (Bapna, Goes and Gupta 2003). Some research has investigated the 

impact of a consumer’s background, such as experience, on their online-auction behavior 

(e.g., see Wilcox 2000). 

Zwick and twelve other scholars outline key auction concepts and empirical findings, 

identify some areas for future research, highlight that Internet auctions can lead to new 

principles in marketing theory and practice, and report that the marketing literature on 

auctions is “sparse” (Chakravarti et al. 2002). They urge marketers to undertake more 

research in auctions. Online auction feedback 

122

Online auction feedback, or reputation and rating information, is a type of word-of-web that 

auction participants provide and pass along to other auction-community members in order to 

assist them in reducing uncertainty associated with their auction-related decision making, 

such as whether to place a bid, or how to bid. It can be an important means for avoiding 

Akerlof’s (1970) “market for lemons” and creating a healthy market which includes products 

available at a variety of quality levels and associated prices, for example, where buyers may 

be willing to pay a premium for products sold by sellers with superior reputations (Resnick et 

al. 2000). 

Feedback for an individual auction is called a review on eBay. It consists of a details about 

the auction, including: a comment/statement that must be classified as either positive, 

neutral, or negative, the name of the eBay member who left the comment as well as their 

feedback rating, a series of (optional) “Response” statements by the individual to the 

comment as well as “Follow-up” statements by the eBay member who provided the (initial) 

comment, the date and time on which the comment was posted, the auction item number to 

which the comment pertains (which contains an active link to the auction-listing for 30-60 

days after the end of the auction), and indication whether the member being reviewed was 

the seller or the buyer in the auction transaction, which is indicated in the right most column 

with an “S” or a “B,” respectively. Each positive comment begins with the word “Praise.” 

Each neutral comment begins with the word “Neutral.” And, each negative comment begins 

with the word “Complaint.” The remainder of the comment may consist of up to eighty (80) 

characters of the reviewers choice (see Figures 2a-c). 

FIGURE 2a about here 

FIGURE 2b about here 

FIGURE 2c about here 

123

An auction participants’ feedback rating is a quantitative score that is based on the feedback 

reviews received from other community members. Ebay suggests that it be treated as “a 

measure of an eBay user's reputation in the community.” A person’s feedback rating can be 

computed using the following procedures: a) +1 point is tallied for each member who has left 

a positive comment for that person; the sum of these tallies is commonly referred to as the 

(number of unique) “positives” and it is the number of positives listed in a person’s Feedback 

Profile as being “from unique users” (e.g., see Figure 1 - IBM’s number of unique positives 

is 2453); b) -1 point is tallied for each member who has left a negative comment for that 

person; the sum of these tallies is commonly referred to as the (number of unique) 

“negatives” and it is the number of negatives listed in a person’s Feedback Profile as being 

“from unique users” (e.g., see Figure 1 - IBM’s number of unique negatives is 106); and c) 

the (number of unique) negatives are subtracted from the (number of unique) positives to 

yield the feedback rating (e.g., see Figure 1 -IBM’s feedback rating is 2347 1 

= 2453 - 106). 

Research into the impact of online-auction feedback on consumer/bidder behavior has 

considered only quantitative aggregate reputation measures, such as the feedback rating, the 

number of unique positives and the number of unique negatives. Evidence indicates that the 

feedback rating does not fully explain how bidders use feedback information, and that 

bidders place more weight on (the number of unique) negatives than on (the number of 

unique) positives (Ba and Pavlou 2002, Lucking-Reilly 2000, Melnik and Alm 2002). 

Assuming that negatives and positives communicate potential losses and gains for a future 

transaction, respectively, then these results are consistent with the asymmetric impact of 

gains and losses in accordance with prospect theory (Kahneman and Tversky 1981, Tversky 

and Kahneman 1992). 

124

1 

Note that eBay actually indicated a feedback rating of 2349 for IBM while reporting 2453 

positives and 106 

negatives; however, this is, obviously, in error. EBay acknowledges this problem in their 

discussion boards: for 

example, a May 26, 2003 post by an eBay staff member states, “Hi Folks, We are aware that 

some members' 

feedback scores and/or summaries are not showing correct numbers. We are currently 

investigating the situation 

and hope to have it resolved soon. Thank you for your patience in this matter as we work to 

correct this problem. 

Daphne Community Development 

A main limitation of these inquiries is that they do not consider many other Feedback Forum 

elements/details, such as individual-auction comments, which contain “much (service 

performance) information about sellers,” (Ba and Pavlou 2002, page 256). Research which 

delves below the surface of the aggregate feedback measures is important because it can 

provide greater insight into: a) how consumers process and use online auction word-of-web 

information, b) which feedback elements impact consumer/bidder perceptions and behavior, 

c) how practitioners should make online-auction reputation- or service-related decisions 

(e.g., being less concerned about not receiving positive feedback, but doing everything 

possible to minimize the likelihood of receiving negative feedback), and d) the role of word- 

of-web rating-and-review information in online auctions in order to replicate or transfer its 

effects to other online (or offline) contexts or applications. 

STUDY ONE: RATING-AND-REVIEW WOW 

125

The objective of study one was to identify a comprehensive set of websites containing rating- 

and-review (RR) information in order to identify important characteristics of the rating-and- 

review information available online. 

We selected a sample of websites to inspect with Google, the online search engine, using 

relevant (search) phrases such as “rating website,” and “product ratings.” We found eighty- 

six 

(86) websites that provided rating-and-review information. At the time of the search, it was 

likely that other websites containing rating-and-review information existed. However, the 

sample was viewed as being reasonably comprehensive given that we perused as many as 

ninety-six search-engine page results for some of the (search) terms applied. A complete 

listing of the websites that were analyzed is in the appendix. 

The types of products which were rated and reviewed among the websites were extremely 

diverse. Reporting on the full range of the items would be tedious given that the sample 

websites included those of organizations such as Consumer Reports and epinions, which 

provide rating and review information for hundreds of products. We would not be surprised 

to learn that online ratings and reviews exist for just about any object of interest. Some of 

the products, which could, perhaps, be considered as less than typical, were: table hockey 

equipment, carbohydrate food items, magic tricks, spy software, charitable organizations and 

professors. Rating-and-review word-of-web characteristics 

Our analysis consisted of inspecting the information listed and the processes associated with 

providing or using the rating-and-review information on each site. This yielded eleven 

primary characteristics that can be used to classify RR WOW -- see Table 1. 


126

The sources which provided the rating and review information varied across websites. For 

example, experts provided their assessment of automobiles at Edmunds.com; consumers 

listed their ratings at Amazon.com; and both buyers and sellers detailed their opinions at 

eBay. The target object, that which was rated and reviewed by a source, was either a product 

provider, such as a seller, retailer or manufacturer, a product, such as a good or a service, or a 

customer/buyer. 

Direction is determined by which “sides”/participants in exchange-related behavior provide 

rating and review information. In general, direction is n-way, where n represents the number 

of “sides” who can provide rating and review information. Most RR WOW is unidirectional 

(i.e., one-way), where one party evaluates some aspect of another party; for example (and 

typically), consumers’ evaluations of retailers (consumer � retailer) is available at websites 

such as bizrate and Zagat. Bidirectional (i.e., two-way) RR WOW was observed only at 

online auction sites, such as eBay, where buyers evaluated sellers and sellers evaluated 

buyers (buyer seller). It could be argued, perhaps, that epinions or Amazon list 

bidirectional RR 

WOW, in that one can find a buyer’s or consumer’s rating of products as well as assessments 

of the buyer/consumer’s rating by other buyers/consumers. 

By definition, rating and review information includes at least one quantitative piece of 

information, a rating. However, some sites provide a breadth of statistics, such as an overall 

rating and a set of attribute ratings (e.g., see bizrate), or an overall rating and a frequency 

distribution (e.g., see eBay). Qualitative review information may also be available. The 

amount of review prose associated with a rating varies. For example, the maximum amount 

allowed at ratemyprofessor.com is 255 characters, while there does not appear to be a limit at 

epinions. Also, the qualitative review information presented may be constituted of raters’ 

127

excerpts that were selected and fused together by those who manage the site. For example, 

Zagat creates restaurant reviews that are based on bits and pieces of opinions provided by 

some of the raters. Some sites present only a sample of raters’ comments (e.g., bizrate). 

Sites also varied by the manner in which RR information was presented and could be 

customized by users. For example, eBay shows summary information by time, and 

gamereview.com enables users to view information by rating level. Some sites integrated 

RR information into the buying process, for example, at CNET, one can quickly access such 

information while reviewing a list of sellers and their prices (and vice-versa). 

Contacting the person who provided the information was available at some (57%) of the 

sites. This facilitates the opportunity for one to probe and learn more from the source of the 

RR information. The length of time in which all RR-related information was available for 

review ranged from a limited period of time to – so it appeared – permanently. For example, 

at Amazon.com, all product reviews appear to remain available forever. Whereas, on eBay, 

some RR-related information elements are available for only 30-60 days. Online auction site 

uniqueness 

Our analysis found that the leading major online-auction sites differ from all other types of 

RR WOW sites. Online auction feedback can be bidirectional in that both buyers and sellers 

may supply feedback about each other. In any RR WOW system, one would expect that 

buyers would be able to provide feedback about sellers. However, in online auction sites, 

such as eBay, sellers can also provide feedback on buyers. 

STUDY TWO: CONSUMER USE OF ONLINE AUCTION RATING-AND-REVIEW 

SYSTEMS 

128

In study one, we learned more about the general makeup of rating-and-review word of web 

and discovered that online-auction RR WOW is unique in that both the buyer and the seller 

can rate each other. The objective of study two is to learn more about how consumers use 

online-auction feedback elements to make decisions. Method 

Eight subjects with experience using online auctions were instructed to inspect auctions for 

products that they would consider obtaining in an online auction. They provided verbal 

protocols of their search. After their experience, subjects were interviewed about their 

utilization of, and perceptions about, feedback information on the auction website. 

eBay was the auction site used in this study as it has been used in a majority of previous 

online auction studies. In addition, with its market share of 80%, many people perceive it as 

“The” online auction website. Results 

Customer perceptions about RR WOW are presented by the type of feedback information 

available on eBay. All reputation information about an online-auction member appears in 

her/his feedback profile page(s). The feedback profile is organized into three main sections: 

Feedback Summary, eBay ID Card (including ID History and Feedback About Others), and 

auction feedback reviews (again, see Figure 1 for an example). (Overall) feedback rating 

In general, the feedback rating was important to subjects from the perspective of being a 

buyer/bidder. For example, subjects said “The total feedback score is important,” “since 

people on eBay don’t know who you are, the only way people know if you’re honest and 

dishonest is through the feedback system.” They also indicated that the feedback rating was 

important from the perspective of perhaps becoming a seller: “My feedback score is 

instrumental in helping me to eventually become a seller,” and “I’d work hard to get enough 

positive feedback so that I can use (i.e., sell with) the ‘Buy it Now’ feature.” 

129

Consumers believed that sellers with relatively high scores were more professional, 

experienced, and trustworthy: “People with lots of feedback are trustworthy because it’s like 

their business,” “The bigger the rating the better,” “If a seller has several hundred feedback 

points or more, then selling things over eBay must be their full-time job,” “A seller who has 

more than several hundred feedback points is more professional than someone with fewer 

points,” “The higher the overall number, the more experienced the seller.” And some 

ascribed economic value to a feedback rating in terms of experience – “It’s worth it to pay a 

little bit more for an item that’s being auctioned by a more experienced seller.” 

Some consumers had a feedback rating threshold-level when assessing a seller: “I’m 

skeptical of sellers that have ratings below 30,” “I’d be uncomfortable bidding more than $25 

for an item that’s being sold by an inexperienced buyer; one with an overall rating below 30” 

and “Fifty or more is a good amount.” 

Every consumer stated that they would not place a bid on an item based solely on the 

feedback rating; they said they would view other feedback information that is available on 

the Feedback Profile pages, such as the individual-auction reviews. Feedback Summary 

Though consumers highlighted the importance of a high feedback rating, they were quick to 

focus on the number of unique negative feedback comments when viewing a seller’s total 

number of unique positives, neutrals and negatives. Consumers’ believed that “It’s better to 

have no negative feedback than to have lots of positive feedback,” and “Even if someone has 

a rating of 0, no negative feedback is an important factor.” This was consistent with research 

which found that a negative comment reduces a seller’s reputation more than a positive 

comment increases it (Ba and Pavlou 2002, Lucking-Reilly 2000, Melnik and Alm 2002 

2000). 

130

However, there is more to understand about negative feedback as consumers also suggested 

that there was a “to be comfortable” tolerance for some maximum number or percentage of 

negative comments: “If someone had a 30 point rating, I would feel comfortable if they had 

less than 3 negatives. If a rating of 10, then 1 negative tops, rating of 50 then 4-5 negatives 

would be ok, rating of 1000+, then up to 20 negatives,” “A seller is bad if they have 1000 

positives and more than 50 negatives,” “If a seller has 10 feedback points and 1 or more 

negatives, then they are a bad seller,” “If a seller has 100 feedback points and 5 or more 

negatives, then they are a bad seller,” and “Once you get 20 or more negative feedbacks as a 

seller, then you’re doing something wrong. A 10% rule holds true for up to 200 transactions 

and then anything above that the 20-only rule kicks in.” It would be interesting to explore 

farther the relationship between perceptions of a seller and percentages or absolute amount of 

negative word-of-web feedback. 

The vast majority of consumers devoted very little attention to the total number of neutral 

feedback comments. However, some believed that it was important to review the comments 

associated with neutral feedback in order to search for possible negative information. In 

general, the aggregate number of neutral comments had no effect on a consumer’s 

perceptions of a seller. 

eBay ID Card 

The primary purpose of the ID card appears to be reporting on the aggregate number of 

feedback comments received over various lengths of time. Not all consumers dedicated time 

and effort toward viewing the eBay ID card; those who did, indicated clearly that feedback 

received more recently was more important: “If a seller has no recent auction activity (i.e., 

any feedback received within the past 6 months), then they are less trustworthy,” “One 

negative feedback within the past 7 days affects my opinion of a buyer because that one 

131

negative feedback could be me,” and “Recent negative ratings are more detrimental than 

older negative ratings.” The first statement might suggest that consumers place a higher 

value on sellers that are consistently involved in the eBay community over time. 

Some attention was devoted to other aspects of the ID card. One person indicated that 

“geography” influences a bidding decision in that “It’s a pain to buy from people who are 

outside the United States.” None of the consumers had actively viewed the “feedback left for 

others” when evaluating feedback. However, one person commented “One good turn 

deserves another. If people in general give a seller positive feedback, then that seller should 

be giving praise too,” when asked what would be expected in the “feedback left for others” 

from a seller that had all positive feedbacks. Interestingly, Resnick et al. (2000) suggests that 

reciprocity might decrease the effectiveness of eBay’s Feedback Forum. 

Auction Feedback Reviews 

Auction feedback review information is available at the individual-auction level for a 

member (again, see Figures 2a-c for examples), and reports, most importantly, the feedback 

comment (and follow-up) provided by the “other” exchange party from that auction, and any 

replies by the member to the feedback comment. Up to twenty five feedback reviews are 

presented on a webpage unless one indicates the other possible viewing-alternatives of 50, 

100, or 200 reviews per webpage. Until recently, an option for viewing up to 500 feedback 

reviews per webpage was offered by eBay. (One main effect of the viewing-alternatives 

option is the number of webpages over which feedback reviews are presented, of course. For 

example, the auction feedback review information for a person who has received 400 

reviews, could be viewed over two pages containing 200 reviews each, or sixteen pages 

containing 25 reviews each. This appears to matter as consumers searching for negative 

comments rarely looked through more than 2 pages of feedback reviews.) 

132

Consumers, again, were quick to focus on Complaints – a negative comment for an auction. 

Interestingly, consumers processed the “detail” of a negative comment and not all complaints 

“appeared to be created equal.” Consistent with beliefs stated about the ID Card, the (time) 

order in which feedback comments appear on a webpage may have an impact, e.g., “It’s not 

good when the feedback comments start with a complaint” and “ It’s bad when someone’s 

most recent comments are negative.” In addition, the comment could have varying impact 

depending on the service-related attributes mentioned: “I’m unhappy when I see that 

someone has a negative comment that states they’re a slow responder to emails, such as ‘took 

too long to receive item’ ” and “It’s worse to have a negative comment about the product, 

like ‘product not as advertised,’ than it is to have a negative comment about slow 

communications, like ‘not responsive to emails’.” Some consumers were looking for very 

specific service performance-related information in the comments, e.g., “Looking for product 

comments, process/shipping, and communication.” Many said they would avoid a seller who 

has feedback comments that are threatening or too personal.” Finally, consumers could 

overlook negative feedback if it was atypical, “If only one person leaves a complaint to a 

seller that has a high feedback score, then that complaint seems less valid.” 

Consumers also processed the comments associated with Praise – positive feedback for an 

auction. Some consumers believed that positive comments indicated a high degree of 

customer satisfaction: “People who leave positive feedback must be very satisfied” and “If a 

seller has all praises, then everyone is happy with them.” They valued generalizations, such 

as “A+++, great eBayer” as well as specific attribute-level details such as “super fast 

shipping.” However, some consumers believed that one should not necessarily take positive 

feedback at face value as “reading between the lines” could reveal important information: “A 

133

idder should read the praises too because it could be a veiled complaint,” and “It would be 

weird if someone left positive feedback, but didn’t leave a comment.” 

Consumers also considered a member’s response to a negative feedback comment. 

Interestingly, consumers suggest that an opportunity exists to, in spirit, turn a negative into a 

positive or, as in service recovery (Maxham 2001, Smith and Bolton 1998), reduce the 

deleterious effect of a negative comment situation: “When someone leaves a response to a 

complaint, it shows that they care,” “When someone leaves a response, it shows they want to 

receive good feedback all of the time” and “It’s good to say that you’re upset in your 

response if you’ve got a few negatives and a high feedback score.” On the other hand, 

consumers suggested that not all types of responses are effective: “Responses that do not 

address complaints directly basically admit that the complaint is true” and “Spiteful or 

retaliatory responses to negative feedback are not good.” 

eBay also indicates in each feedback review whether a member was the buyer or the seller. 

Although a minority of the participants regularly looked at this information when reviewing 

feedback details, most believed that, for a seller, a Praise comment was worth more if it was 

earned through selling than through buying. However, one participant found that there was 

value in seeing that a seller was also good buyer, saying “It’s important that a seller receives 

positive comments when they are a buyer.” Basically, this consumer felt that a well-rounded 

eBay user who experienced both aspects of transactions would be more in tune with the 

needs of buyers and, as a result, would deliver better products and service. A few consumer 

wishes 

Consumer participants in this study would like eBay to enable additional search-features and 

completed-auction information to facilitate their assessment of sellers. Consumers were 

quick to peruse negative comments and therefore would like eBay to facilitate “Searches for 

134

an easy way to see only negative comments,” or in general, to sort a member’s feedback by 

feedback comment type (i.e., positive, neutral or negative). They also expressed a desire to 

see more detailed feedback information, perhaps something along the lines of attribute 

ratings, e.g., 

“Want to see serious sentences as to why someone is satisfied when they leave a positive 

comment and why someone is dissatisfied when they leave a negative comment.” 

In addition, consumers would like “to see a link to the auction page of an item related to any 

negative feedback even if it’s past the time when those links are kept active,” basically, to 

better determine the fairness and honesty of a seller. For example, sellers may reply to 

negative feedback by stating “view the auction details” as a means for validating their 

integrity. However, when the link to the auction listing is no longer active -- eBay’s practice 

is to list a completed auction for about 30-60 days -- then one can no longer view the 

auction-listing details and more fully validate the trustworthiness of a seller. Perhaps some 

potential bidders give sellers some benefit of the doubt when they see “trust-supporting” 

statements that can not be confirmed at a current time through no fault of the seller. 

SUMMARY AND FUTURE RESEARCH 

An objective of this research was to extend our understanding of online auctions, which 

leading marketing academics and professionals have highlighted as the most important 

Internet application and a research priority. Our investigation focused, ultimately, on eBay’s 

Feedback Forum, which has been considered the key ingredient in the success of the most 

profitable and leading online retailer to date. 

Our approach included defining the concept of word-of-web and its various classes, which 

we used to provide structure to our investigation, and then implementing two studies. The 

135

first study focused on one class of WOW to which online auction feedback belong, rating- 

and-review, and identifying a set of characteristics on which RR WOW may vary. The 

second study focused on consumer processing of online auction RR WOW, specifically, 

eBay’s Feedback Forum. 

Study one revealed that RR WOW for online auctions differs from that of all other online 

sources on one characteristic, which heretofore has not been an issue in retailing, that of 

direction. Online auction sites such as eBay, Amazon-Auctions and Yahoo!-Auctions each 

provided bidirectional RR WOW, whereas all other sites provided unidirectional RR WOW. 

We believe that retailers, among others, need to learn more about bidirectional RR WOW. 

Study two identified several important areas for new exploration into online auction RR 

WOW. Previous research looked at the relationship between the (overall) feedback rating 

and various online auction behaviors/results (e.g., final bid amount). However, our research 

shows that, in addition to the feedback rating, consumers inspect and consider many other 

pieces of information in online auction RR WOW, such as the individual-auction level 

feedback reviews. 

This suggests significant opportunity for future research. In general, future research can 

explore further the impact of various elements of online auction RR WOW on a consumer’s 

perception of a retailer/seller and a consumer’s decision making behavior (e.g., whether to 

place a bid, setting a bid-amount limit, whether to provide feedback). For example, though 

online auction providers (e.g., eBay) assign the same value to each type of auction rating 

(e.g., +1 for Praise, 0 for Neutral, -1 for Complaint), do consumers assign the same value to 

each type of rating for all auctions? Are all positively-rated auctions (experiences) accorded 

the same weight? Are all negatively-rated auctions (experiences) viewed in the same way? 

Study two suggests that this may not be the case given that consumers consider the 

136

order/time in which feedback was received and the detail of the qualitative feedback review 

comments of online-auction RR WOW. 

Also, future research could assess which service-performance attributes are most important 

to online-auction consumers/bidders, determine how seller’s should “respond” to negative 

comments and manage service recovery or brand building, or consider the impact of a 

consumer’s ability to sort comments by their type, positive, neutral or negative (such sorting 

is indeed feasible at http://www.vrane.com). 

At a broader level, detailed research into other types of word-of-web that are commonly used 

online by retailers, such as e-newsletters or email, could prove very profitable. In fact, at the 

present time, learning more about these two particular types of WOW could be extremely 

important given the prevalence and negative perceptions of Spam-email. 

REFERENCES 

Adler, Jerry. 2002. The eBay way of life. Newsweek. June 17. 50-60. 

Akerlof, George. 1970. The Market For ‘Lemons’: Quality Uncertainty and the Market 

Mechanism. Quarterly Journal of Economics. 85:3. 488-500. 

Ba, Sulin and Paul A. Pavlou. 2002. Evidence of The Effect of Trust Building Technology in 

Electronic Markets: Price Premiums and Buyer Behavior. MIS Quarterly. 26:3. 243-268. 

Bapna, Ravi, Paulo Goes and Alok Gupta. 2003. Analysis and Design of Business-to- 

Consumer Online Auctions. Management Science. 49:1. 85-101. 

Bickart, Barbara, and Robert M. Schindler. 2001. Internet Forums As Influential Sources Of 

Consumer Information. Journal of Interactive Marketing. 15:3. 31-40. 

Chakravarti, Dipankar, Eric Greenleaf, Atanu Sinha, Amar Cheema, James C. Cox, Daniel 

Friedman, Teck H. Ho, R. Mark Isaac, Andrew A. Mitchell, Amnon Rapoport, Michael H. 

137

Rothkopf, Joydeep Srivastava, Rami Zwick. 2002. Auctions: Research Opportunities in 

Marketing. Marketing Letters. 13:3. 281-296. 

CNET News. 2002. eBay: Last man standing. CNET News. April 20. online at 

http://news.com.com/2009-1017-887630.html. 

Cristol, Hope. 2002. News in the Digital Age. The Futurist. 36:5 (Sep/Oct). 8-9. 

eBay. 2003. eBay Inc. Announces First Quarter 2003 Financial Results. Online at 

http://www.shareholder.com/ebay/news/20030422-107123.htm. Gelb, Betsy D. and Suresh 

Sundaram. 2002. Adapting to “word of mouse.” Business Horizons. 

45:4. 21-25. 

Gilkeson, James H. and Kristy Reynolds. 2003. Determinant of Internet Auction Success and 

Closing Price: An Exploratory Study. Psychology and Marketing. 20:6. 537-566. 

Gilly, Mary C., John Graham, Mary Wolfinbarger and Laura Yale. 1998. A Dyadic Study of 

Interpersonal Information Search. Journal of the Academy of Marketing Sciences. 26:2. 83- 

100. 

Godin, Seth. 2001. Unleashing the Ideavirus. Hyperion. 

Grant, Elaine X. 2002. What Makes Ebay Invincible. E-Commerce Times. March 5. Online at 

http://www.ecommercetimes.com/perl/story/16546.html. 

Hagel, John III, and Arthur G. Armstrong. 1997. Net gain: expanding markets through 

virtual communities. Boston, MA: Harvard Business School Press. 

Hansen, Ward. 2000. Principles of Internet Marketing. Cincinnati, OH: South-Western 

College Publishing. 

Hof, Robert D. 2001. The People’s Company. BusinessWeek E.Biz. December 3. Available 

online at http://www.businessweek.com:/print/magazine/content/01_49/b3760601.htm?eb. 

138

Kahneman, Daniel and Amos Tversky. 1979. Prospect Theory: An analysis of decision under 

risk. Econometrica, 47:2, 263-291 

Katz, Michael. 2002. E-Newsletters That Work. Electronic book available online at 

http://www.enewsletterbook.com/. 

Kambil, Ajit, and Eric van Heck. 2002. Making Markets. Boston, MA: Harvard Business 

School Press. 

Kozinets, Robert V. 2002. The Field Behind The Screen: Using Netnography For Marketing 

Research in Online Communities. Journal of Marketing Research, 29:1. 20-38. 

Kuehl, Claudia. 1999. New World of Web reviews. Internet World. 5:34 (December 1). 52- 

54. 

Levy, Steven. 2002. Living in the Blog-osphere. Newsweek. 140:9 (August 26). 42-45. 

Lucking-Reiley, David. 2000. Auctions On The Internet: What’s Being Auctioned And How. 

The Journal of Industrial Economics. 48:3. 227-252. 

Lucking-Reiley, David, Doug Bryan, Naghi Prasad, and Daniel Reeves. 2000. Pennies from 

eBay: the Determinants of Price in Online Auctions. Available online at 

http://www.vanderbilt.edu/econ/reiley/papers/PenniesFromEBay.pdf. 

Maxham III, James G. 2001. Service Recovery’s Influence on Consumer Satisfaction, 

Positive Word of Mouth and Purchase Intentions. Journal of Business Research 54:1. 11-24. 

Maxham III, James G. and Richard G. Netemeyer. 2002. A Longitudinal Study of 

complaining Customer’s Evaluations of Multiple Service Failures and Recovery Efforts. 

Journal of Marketing. 66:4. 57-71. 

McAfee, R. Preston., and John McMillan. 1987. Auctions and Bidding. Journal of 

Economics Literature. 25:X. 699-738. Melnik, Mikhail I. and James Alm. 2002. Does a 

Seller’s Ecommerce Reputation Matter? Evidence from Ebay Auctions. The Journal of 

139

Industrial Economics. 50:3. 337-349. Milgrom, Paul R. 1989. Auctions and Bidding: A 

Primer. Journal of Economic Perspectives. 

3:X. 3-22. Milgrom, Paul R. and Robert J. Weber. 1982. A Theory of Auctions and 

Competitive Bidding. 

Econometrica. 50:5. 1089-1122. Ong, Walter J., S.J. 1982. Orality and Literacy. London: 

Methuen. Ottaway, Thomas A., Carol L. Bruneau and Gerald E. Evans. 2003. The Impact of 

Auction Item 

Image and Buyer/Seller Feedback Rating on Electronic Auctions. The Journal of Computer 

Information Systems. 43:3. 56-60. 

Resnick, Paul, Richard Zeckhauser, Eric Friedman and Ko Kuwabara. 2000. Reputation 

Systems. Communications of the ACM. 43:12. 45-48. 

Rheingold, Howard. 2000. The Virtual Community: Homesteading on the Electronic 

Frontier. Cambridge, MA: MIT Press. 

Richins, Marsha. 1983. Negative Word of Mouth by Dissatisfied Consumers: A Pilot Study. 

Journal of Marketing, 47:1. 68. 

Scientific American. 2001. Patent Pamphleteer. Scientific American. 285:6 (December). 33. 

Smith, Amy K. And Ruth Bolton 1998. An Experimental Investigation of Service Failure and 

Recovery: Paradox or Peril? Journal of Service Research, 1(August). 65-81. 

Srinivasan, Srini, Rolph Anderson and Kishore Ponnavolu Customer Loyalty in E- 

Commerce: An Exploration of Its Antecedents and Consequences. Journal of Retailing. 

78:1.41-50. 

Stern, Barbara 1994. A Revised Communication Model for Advertising: Multiple 

Dimensions of the Source, the Message and the Recipient. Journal of Advertising. 23:2. 5- 

15. 

140

Tversky, Amos and Daniel Kahneman. 1992. Advances in Prospect Theory: Cumulative 

Representation of Uncertainty. Journal of Risk and Uncertainty. 5, 297-323. 

US Census Bureau. 2002. Retail 1Q, 2003 E-commerce Report. May 30. Available online at 

http://www.census.gov/mrts/www/current.html 

Wilcox, Ronald T. 2000. Experts and Amateurs: The Role of Experience in Internet 

Auctions. Marketing Letters. 11:4. 363-374. 

APPENDIX - Websites with Rating-and-Review Information 

http://activebuyersguide.com/ 

http://altavista.dealtime.com/main/1,3241,,00.html?header= 

http://ask.dealtime.com/dealtime2000/Reviews/product/read_product/1,7235,20165,00.ht 

ml http://borntolove.com/cgi-bin/productreviews/ http://cg.superpages.com/cgibin/php/index.html 

http://epguides.com/Seinfeld/ 

http://home.labvelocity.com/researchlink/index.jhtml 

http://infocenter.cramsession.com/TechLibrary/default.asp?CatID=395 

http://reviews.cnet.com/reviews/0-9870989.html 

http://shopping.yahoo.com/reviews/productopia/ 

http://sportsbay.com/productratings.html http://theaquarians.net/user_reviews.htm 

http://216.234.172.45/cgi-bin/starsreview/stars.pl http://tv.zap2it.com/news/ratings/ 

http://wsj.consumersearch.com/electronics/cd_players/reviews.html 

http://www.acnielsenedi.com/bonews/bonewsframes.html 

http://www.activebuyersguide.com/ http://www.atkinsfriends.com/reviews/ 

http://www.audioreview.com/reviewscrx.aspx 

http://www.babyplus.com/science/product_ratings.html 

http://www.bikemagic.com/review/review.asp http://www.bikergirl.net/products/ 

http://www.biowire.com/bw_jsp/product_reviews_top.jsp 

http://www.bodypaincentral.com/ http://www.carreview.com/reviewscrx.aspx 

http://www.charitynavigator.org/ http://www.circuitcity.com 

http://www.computingreview.com/reviewscrx.aspx http://www.consumerguide.com/ 

http://www.consumerreports.org/main/home.jsp http://www.consumerreview.com/ 

http://www.consumersearch.com/www/ http://www.corvetteforum.com/reviews/ 

http://www.creativepro.com/reviews/swproduct http://www.edmunds.com/ 

http://www.epinions.com/ http://www.excelsis.com/1.0/section.php?sectionid=17 

http://www.gamerankings.com/itemrankings/latestadditions.asp?platform=15 

http://www.gamespot.com http://www.gamezone.com/reviews/reviews.htm 

http://www.golfreview.com/reviewscrx.aspx http://www.golftestusa.com/ 

http://www.gsj.com/prodrateChoice.asp 

http://www.hairlosshelp.com/html/productratings.cfm 

http://www.headshaver.org/reviews/ http://www.heavyweights.net/productrating.asp 

http://www.internet-magazine.com/reviews/index.asp 

http://www.isedb.com/index.php?t=reviews&id=1067 http://www.kbb.com/ 

141

http://www.laxshopper.com/index.htm?/equipment/ratingslistings.htm~laxmain 

http://www.momsjewelrybox.com/Our-Ratings.htm http://www.mtbreview.com/reviews/ 

http://www.nextag.com/serv/main/buyer/help/faq.jsp?a=b 

http://www.nuws.net/hardware/ 

http://www.outdoorreview.com/jump%5Epage%5Eflycrx.aspx 

http://www.paddling.net/Reviews/Kayaks.phtml 

http://www.pcgamereview.com/whatsnew/whatsnew.shtml 

http://www.pcphotoreview.com/ http://www.pcri.net/ratings.htm 

http://www.pcworld.com/reviews/index/0,00.asp http://www.planetfeedback.com/ 

http://www.pricegrabber.com/about.php/about=us/ut=183c0ac1b09d7959 

http://www.rateitall.com/ http://www.ratemyprofessor.com/ 

http://www.ratemyteacher.com/ http://www.ratings.net/ 

http://www.responsibleshopper.org/ http://www.reviewcentre.com/ 

http://www.riversmallies.com/reviews.html 

http://www.roadbikereview.com/reviewscrx.aspx 

http://www.scatmat.com/Tools/ProductRatings/AddProduct/ 

http://www.scrapbookaddict.com/reviews/ http://www.sillymagician.com/magic.htm 

http://www.sleestak.net/reviews/ http://www.spy-software-products.com/ 

http://www.stigaonline.com/rankings.shtml 

http://www.substance.com/discuss/rate/reviews/0,11191,228622,00.html 

http://www.sysopt.com/userreviews/ 

http://www.systemid.com/products/product_walkthrough.asp 

http://www.thinnerthoughts.com/reviews.html 

http://www.tvtome.com/tvtome/servlet/GuidePageServlet/showid-112/epid-2241/ 

http://www.uk.ciao.com/rating/The_Best_Films_on_DVD.html 

http://www.uscomputer.net/reviews/reviews.htm 

http://www.virtualratings.com/frames/productratingsiss.html http://www.zagat.com/ 

http://www.zotz.com/nissan/sentra_product_ratings.htm 

Figure 1 - Feedback Profile page excerpt of eBay member IBM Figure 2a - Example of 

Positive Comment (left by kt_live for IBM) 

Figure 2b - Example of Neutral Comment (left by madcoder42 for IBM, with a Response 

by IBM to the comment) 

Figure 2c - Example of Negative Comment (left by rtadder for IBM, with a Response by 

IBM, and a Follow-up statement by rtadder) Table 1 Characteristics Of Rating-and- 

Review Word of Web 

? Source The source of the information, such as buyer, employee, 

analyst/expert, or seller. 

? Target Object The target object of the RR information, such as brand, 

product, manufacturer, or retailer. 

142

? Direction Direction is determined by the participants and “sides” in 

exchange-related behavior who provide RR information. In 

unidirectional or one-way RR WOW, a single participant, 

such as a consumer, provides information about a target 

object, such as a seller/retailer. In bidirectional or two-way 

RR WOW, the buyer provides feedback about the seller 

and the seller provides feedback about the buyer. 

? Statistics The rating statistics available, such as an overall rating, 

frequency distributions, and attribute ratings. 

? Qualitative Information The extent to which one may provide qualitative 

information, and the availability of relevant qualitative 

information, such as prose that explains a rating. 

? Selection The extent to which the RR information may not be 

presented in its original form, such as by selecting excerpts 

or censoring. 

? Quantity/Sampling The quantity of reviewer comments available, such as from 

? Presentation Format and 

Customization 

all reviewers, or a sample/proportion of them. 

The format in which information are presented including 

order, utilization of graphics and the like; and the extent to 

which the format may be customized, such as sorting 

retailers from best to worst, or listing only “negative” 

comments. 

? Transaction Integration The extent to which the information is integrated into or 

accessible during the transaction process, such as 

presenting RR WOW prominently on a webpage that 

presents, describes, and enables purchase of a product. 

? Probative Nature The ability to interact with the source of the RR WOW, 

such as via email. 

? Permanence The extent to which RR WOW remain available, such as in 

detail, or in time. 

143

144

Physical Behaviour in Stores and Spatial Configuration: Two Exploratory Studies 

Gaël Bonnin, EDHEC Business School 

Atmospheric perspective (Kotler, 1973; Russel and Mehrabian) has a great influence on 

empirical works studying physical environment-shoppers relationships. Nonetheless, recent 

works have mentioned certain limits of this perspective. Drawing from propositions of Titus and 

Everett (1995) and from the frame of analysis of European psycho-sociology of environment, the 

aim of this paper is to show the importance of physical behaviours in the creation of shopping 

experience and the influence of spatial configuration on these behaviours. 

Research has emphasized the role of the physical environment as a means to increase the 

performance of retail chains (Turley and Chebat, 2002; Babin and Attaway, 2000; Bitner, 1992; 

Kotler, 1973). However, recent works have illustrated that the use of atmospherics tools is not 

always profitable (Kozinets and al, 2002). 

It seems then of great importance to understand more clearly the relationship between consumers 

and the physical environment in which they behave. 

The atmospherics model (Russel and Merhabian, 1976; Kotler, 1973) is a great step towards this 

understanding. Empirical works drawn upon this model have shown that physical environment 

cues can affect people‘s internal states (mainly affective states) and their behaviours (either 

physical or purchase related) (Chebat and al, 2001; Chebat and Robicheaux, 2001; Lam, 2001; 

Turley and Milliman, 2000). 

Yet, recent works have highlighted certain limits of this framework (Aubert-Gamet and Cova, 

1999; Everett and al, 1994; Sherry, 1998; Stokols, 1978). We mention three of them. 

145

Firstly, the consumer is supposed to be a passive stakeholder in the physical environment- 

individual relationship: only the influence of environment on consumers is studied whereas the 

study of people’s influence on environment is neglected. 

Secondly, research is focused on sense-stimulating cues of the environment. Other dimensions, 

such as spatial configuration, though sometimes mentioned, does not really fit in the perspective 

of this model. 

Thirdly, people’s experience of the environment is supposed to be based on internal states, whilst 

physical behaviour is supposed to be a consequence and not an endogenous variable. But, a few 

works have proposed that behaviour may be an important means for the consumer to create its 

experience (Sherry, 1998; Fischer, 1997; Moles and Rohmer, 1998; De Certeau, 1980). 

A first attempt to challenge these limits was made by Titus and Everett (1995), who proposed the 

following: 

- Patterns of behaviour (i.e. movement and contact) are linked with types of experience 

consumers are looking for. Two search strategies are distinguished: epistemic strategies and 

hedonic strategies; 

- These patterns of behaviour may be linked with spatial configuration. 

As far as we know, the work of Titus and Everett has not been followed by empirical 

investigations. 

The aim of this paper is therefore twofold: 

Explore the strategies used by consumers in stores; 

Measure the influence of spatial configuration on these strategies. 

The two following studies were carried out to reach these objectives. 

I. Study 1: Behavioural Strategies of Consumers in Stores 

146

I.1. Theoretical Background 

To analyse behavioural strategies, Titus and Everett (1995) drew from the work of Darden and 

Dorsch (1990) who defined shopping strategies as “a plan for the exertion of energy for the 

performance of a shopping behaviour”. Navigation search strategies are behavioural strategies 

and are a smaller subset of all shopping strategies. They “help direct consumer’s effort to locate 

and acquire desired products, information and sensory stimulation within retail shopping 

environments”. For the authors two broad classes of strategies may exist : epistemic and hedonic. 

Epistemic search strategies are designed for the sole purpose of locating and purchasing desired 

products within retail shopping environments. 

Hedonic search strategies are used to enhance the overall quality of the multisensory shopping 

experience and satisfy the shopper’s hedonic pursuit of pleasure. 

These strategies are directly linked with physical behaviours along two dimensions: movement 

and contact. 

Shoppers using hedonic search strategies will proceed more slowly, stop more frequently, pause 

longer and backtrack less often than shoppers using epistemic search strategies; 

Shoppers employing epistemic search strategies will interact less often, and for shorter periods of 

time, with other customers and products than shoppers using hedonic search strategies. 

As far as we know, no empirical works have drawn from Titus and Everett (1995) frame of 

analysis. But, we have identified two European works carried out in non-retail contexts, that 

have drawn from European psycho-sociology of environment (Fischer, 1997; De Certeau, 1980; 

Moles and Rohmer, 1998). The relevance of this perspective lays in the fact that it gives a better 

understanding of physical behaviour sense and of the interaction mechanisms of environment- 

person relationship. 

147

For the psycho-sociology of environment, a key concept in the relationship between people and their 

environment is appropriation. Appropriation is a process by which people exert their influence on the 

physical environment (Fischer 1997; Moles and Rohmer, 1998). Through appropriation people create 

space‘s experience and build attachement to a place (Bones et Secchiaroli, 1995; Moles et Rohmer, 1982; 

De Certeau, 1980; Fischer, 1981). 

Every physical act is an appropriation practice (Moles et Rohmer, 1977), but two elementary 

aspects have been distinguished (Fischer, 1981): visual activity (minimal means of 

appropriation) and exploration (based on movement and contact). 

These appropriation practices, once combined, constitute appropriation strategies. These are ways of 

exploring the physical environment and are formal structures limited in number. Each has an underlying 

signification, referring to the “lived” experience in space. 

Moles and Rohmer (1977), like Titus and Everett (1995), have proposed the existence of two main 

strategies: one functional (epistemic) and one ludic (hedonic). But they proposed to split hedonic 

strategies into active and passive strategies. These differ from each other by the level of involvement of 

the person. Active strategies require high involvement (touching a product) whereas passive strategies 

require lower involvement (just looking at the product). 

The two empirical investigations on the topic have been carried out in a museum and in a 

subway (see Tables 1,2 and 3). These studies have not revealed three strategies but four: the first 

three mentioned above, and another one understood as a denial of the environment and elements 

in it. 

Author Space studied Appropriation practices observed 

Véron and Cultural Walking: point of entrance, space between people and exhibit panels, order of the visit, stoppings 

Levasseur 

(1991) 

exhibition 

Floch (1990) Parisian Actions: absorbed in a book, seating, head lowered; walks straight to a point on the platform (connection); 

subway stops to an exhibition (guitarist); follows the flow, adopt the rhythm of the flow; sits up to look at the 

landscape; listen to the music (walkman), looking nowhere; listens to a conversation; zigzags; does not 

give a helpful hand to a person with heavy luggage at a ticket barrier; is not (like other people) affected by 

repeated billboards; looks at people getting in; speeds up to avoid getting blocked by opposite flows 

Table 1. Appropriation practices in two non-retail spaces 

Véron and 

Levasseur 

Type 1 Type 2 Type 3 Type 4 

Fish 

Walks in the middle of the 

Butterfly 

zigzags, left-right-left– 

148 

Ant 

When he stops, he is close to 

Grasshopper 

Jumps from one point to

(1991) cultural 

exhibit 

Floch (1990) 

subway 

path away from both walls, 

short time (5 to 10 minutes), 

few stopping, goes through 

empty spaces, the itinerary 

looks like a loop, the 

behaviour is always the 

same, does not take care of 

the chronological order 

Continuous 

Follows the flow, ignores 

limits, signs, attraction and 

events, the environment is 

neutralised, same posture, 

same look, same 

concentration over its 

knitting or every other 

activity 

Table 2. Appropriation strategies in two non-retail spaces 

Véron and 

Levasseur 

(1991) cultural 

exhibit 

Floch (1990) 

subway 

right, semi-long time, 

about fifteen stops, avoids 

going through empty 

spaces, behaviour can be 

different from one space to 

another, follows the 

chronological order 

Discontinuous 

Search for and 

appreciation of rhythms 

and role playing offered by 

the environment, tries to 

delimitate, give rhythm, to 

segment, to rediscover 

some places, to oppose 

other places, meticulosity, 

at ease and open to the 

environment 

the exhibit panels, long time, 

numerous stops, does not cross 

empty spaces, walks along a 

wall, same behaviour in the 

two spaces, follows the 

chronological order of the 

exhibition 

Non discontinuous 

Joins, steps over, anticipates 

the obstacles to avoid it, 

minimum, speeds high and 

strong, zigzags, moves 

between, linking, transgression 

another, short time (5 

minutesn), few stops (5 or 6), 

goes through empty spaces, 

same behaviour in the two 

spaces 

Non continuous 

break, pause, interruption 

Type 1 Type 2 Type 3 Type 4 Global interpretation globale 

(and validation methods) 

Fish 

Butterfly 

Ant 

Grass Hopper The behaviors reflect the 

No appropriation, the Controls its visit, Pedagogical Stroll, no motivation attitudes towards the exhibition 

visit is a neutral pleasure, curiosity concern 

concerning the and cultural consumption 

event 

exhibition or its topic (triangulation 

discourses) 

with visitors’ 

Sleep-walking Explorers Pro 

Strollers 

The behaviors reflect the way 

The trip is a neutral The trip is the The trip must be The trip is the people live the moment spent 

event 

opportunity of controlled, efficient opportunity of in the subway, the values of 

entairnment and spare time and passive relaxation subway consumptions (no 

efforts 

mentioned validation method) 

Tableau 3. Interpretations of appropriation strategies in two non-retail spaces 

This short literature review has shown that: 

-people use physical behaviours so as to create space, and therefore shopping, experience; 

-distinguishing two broad strategies, like Titus and Everett (1995), is too limited. 

The empirical investigations mentioned above were carried out in non-retail settings. Do the 

strategies also exist in retail setting? 

The aim of the first study was to answer this question. 

I.2. The Study: 

149

To analyse appropriation strategies in a retail setting, the study took place in the women’s 

clothing area of a hypermarket. Non participant and hidden observation allowed us to collect 

information on 99 shopping trips of different persons. The observation began when an adult 

woman, alone or accompanied, entered the zone. Her route and actions were registered on the 

map of the store. 

Semiotic, and more specifically the semiotic square, was used to analyse the information. 

Following Semprini (1990), we first carried out an analysis of the form of behaviour to 

distinguish strategies and then, on the basis of this analysis, interpret their meanings. 

Every shopping trip observed is a combination of a limited number of appropriation 

practices (see chart 1). 

Figure 1. Elements of a shopping trip 

Movement 

Walking straight 

Turning 

Turning around 

Stopping 

Gestual activity 

Basic gesture 

Touching 

Visual activity 

Basic visual activity 

Return to movement or 

Visual activity 

Basic visual activity 

Location search 

Looking at a particular product 

Looking at a group of products 

To look a product on oneself 

Looking at the label 

Looking and comparing 

different products 

Gestual activity 

Basic gesture 

Touching a product 

Examining a product 

Taking and lifting a product 

Leaning back a product 

S hi i h h lf 

Walking Stopping 

150

Two key moments of a shopping trip are stopping and walking. In each of these 

moments, specific appropriation practices take place. Their combinations define three types of 

stopping and four types of walking (see table 4), with an underlying fundamental and gradual 

opposition between continuity and discontinuity. For example, “walking” means following the 

flow of movement whereas “stopping” is a break in the movement. 

Continuous stopping (cS): the shopper looks at the products, touches a product, look at the label 

Non-discontinuous stopping (ndS): the shopper examines a product, takes and lifts a product, search in the shelf, leaves back a 

product, compares products, takes a product off its hanger. The shoppers does not repeat the actions. 

Discontinuous stopping (dS) : the shopper repeats once or more the following actions: examines a product, takes and lifts a product, 

search in the shelf, leaves back a product, compares products, takes a product off its hanger. 

Continuous walking (cW) : some stoppings (max three), some turns, no turnarounds. 

Non continuous walking (ncW) : several spread out stoppings, several turns, few turnarounds ; 

Non discontinuous walking (ndW) : concentrated stoppings, possibly one or two stoppings outside the concentration zone, few turns, 

turnarounds can be numerous ; 

Discontinuous walking (dW) : numerous stoppings, repeated turns and turnarounds. 

Table 4. Definitions of stoppings and walking types 

The combination of these types of stopping and these types of walking define four 

Types of walking/ 

Types of stoppings 

cW ncW ndW dW 

CS Continuity Non-continuity Non-discontinuity Non-continuity 

NdS Non-discontinuity Non-continuity (one or 

two ndS) 

Non-discontinuity Discontinuity 

Discontinuity 

more ndS) 

(three or 

DS Non-discontinuity Discontinuity Non-discontinuity Discontinuity 

strategies: continuous, discontinuous, non-continuous, non-discontinuous (see table 5). 

Tableau 5. Appropriation strategies as combinations between stopping and walking 

Once the strategies identified, we have to interpret their meanings (see chart 2). 

The continuity strategy is the simplest one. Typically, the route is a straight line 

connecting two points outside the women’s clothing area. For example, the shopper went from 

the central aisle to the bakery or men’s clothing zone. The exploration of the area is minimised. 

The shopper rejects both the environment offer and the product offer of the store. The experience 

created is interpreted as utilitarian in the sense that the aim of the route is to go elsewhere than 

the women’s clothing area. We called this behaviour a “going by” utilitarian shopping trip. 

151

The second strategy is the most elaborated. Movement and stopping are both complex. 

The shopper visit many sub-zones in the women’s clothing area, each time with a great number 

of actions. The exploration is both wide and deep. The shopper interacts highly with the physical 

environment and with the products. The visit may be an opportunity to achieve a precise task, 

possibly planned before coming into the area, but this strategy is not an efficient one. Many of 

the actions can’t be explained by a will to find a product or information on a product as fast as 

possible. On the contrary, they seem only aimed at introducing playfulness in the shopping trip. 

The experience created is a ludic one, with the consumer acting actively on the environment. We 

called this strategy a ludic active shopping trip. 

The third strategy combines complex stopping and simple movement. The shopper 

interacts mainly with one sub-zone, in which nearly all the actions often numerous are 

accomplished. The other sub-zones are visited only to lead the consumer to the interaction sub- 

zone. The exploration is narrow and deep. There are just as many actions as necessary to find the 

way to the products and to look at them and to choose. The final aim is to find a product or 

information on the product as efficiently as possible. The experience created is functional. We 

called this strategy functional utilitarian shopping trip. 

The fourth strategy combines simple stopping and complex movement. The shopper visits 

different sub-zones but with globally very few actions in each of the sub-zones. The exploration 

is wide but superficial. There are interactions with the global environment but very few with the 

products. It’s just as if the shopper wants to linger in the store but without actually involving 

herself in acting upon it. Time seems to be dedicated to relaxation. The experience created is 

ludic, but with the consumer being rather passive. We called this strategy a passive ludic 

shopping trip. 

152

table 3). 

« Going by » functional 

experience 

Minimum exploration 

Continuity 

Functional utilitarian 

experience 

Narrow and deep 

exploration 

Non discontinuity 

Active ludic experience 

Wide and deep 

exploration 

Discontinuity 

Passive ludic experience 

Wide and superficial 

exploration 

Non-continuity 

Chart 2. Meanings of the appropriation strategies observed in a retail setting 

Our interpretation is convergent with other works for the same types of behaviours (see 

Most of studies on environment were drawn from the atmospherics model and implicitly 

assume that consumers are passive and that experience finds its roots in internal affective states. 

Our empirical study supports and complements the work of Titus and Everett (1995). It shows 

that the consumer is active in the creation of experience and that physical behaviours are key 

elements of the space experience. 

A new question is then raised: what dimensions of the physical environment can influence these 

strategies? 

153

From the appropriation perspective (De Certeau, 1980; Fischer, 1997; Moles and Rohmer, 1998), 

the organization of mobility (i.e. spatial configuration) must be studied rather than the senses- 

stimulating cues of the environment. 

II. Study 2: Influence of Spatial Configuration on Behavioural Strategies 

II.1. Theoretical Background 

From the point of view of Titus and Everett (1995), stimulating environment will lead to hedonic 

strategies whereas legible environment will lead to epistemic strategies. But some environmental 

cues are more closely linked with shoppers’ behavioural strategies. 

Shopping environments possessing symmetrical design properties (grid aisle patterns, orthogonal 

path angles) will be perceived as more legible and less stimulating than environments containing 

more asymmetrical design properties (Titus and Everett, 1995). 

For Moles and Rohmer (1982), the basic element of the environment‘s influence on behavioural 

strategies is physical discontinuity introduced in the environment. This takes two forms: 

-The enclosure is the degree to which the zone is open on the outside 

-The micro-events (Moles and Rohmer, 1977) are, inside a zone, variations of a stimulus. It is, 

for example, the end or the beginning of the shelf. 

Combinations between these two elements define three types of space: 

-Functional spaces aim at the efficiency of behaviour within them (Fischer, 1981). They are open 

on the outside and have little and foreseeable micro-events (Moles and Rohmer, 1982). These are 

spaces where nothing happens. They neither raise consumer interest nor make it easy and 

enjoyable for consumers to interact with it. Only necessary actions are accomplished. We should, 

therefore, find more functional utilitarian strategies in these functional spaces than in other 

spaces, everything else remaining constant. 

154

-Active ludic spaces aim at entertaining the shoppers. They are enclosed, with numerous micro- 

events, randomly dispatched. They introduce uncertainty and surprise in the environment. They 

should make it easy and enjoyable for people to behave in and raise the curiosity of shoppers. 

Therefore, there should be more active ludic strategies in these acitvie ludic spaces than in other 

spaces, everything else remaining constant. 

-Passive ludic spaces are open spaces, with micro-events but not at the same level as in active 

ludic spaces. The curiosity of consumers should be moderate leading to superficial actions. 

Therefore, there should be more passive ludic strategies in these passive ludic spaces than in 

other spaces, everything else remaining constant. 

Regarding “going by” functional strategies, we decided not to include them in the study. Indeed, 

these are a very particular case of relationship to shopping environment. The zone is, in this case, 

used as a short-cut between two other points outside the women’s clothing zone. This type of 

behaviour is the result of characteristics present in the environment surrounding, thus outside, 

the women’s clothing area. These are not dependent on spatial configuration of the women’s 

clothing area spatial configuration itself. 

Therefore, our three hypotheses are: 

H1. In an active ludic space, there are more active ludic strategies than in other types of spaces, 

everything else remaining constant 

H2. In a passive ludic space, there are more passive ludic strategies than in other types of spaces, 

everything else remaining constant 

H3. In a functional space, there are more functional utilitarian strategies than in other spaces, 

everything else remaining constant. 

155

To test these hypotheses we selected three spaces (see charts 3 to 5b). These were three women’s 

clothing zone of a hypermarket (Carrefour), and two speciality stores (C&A and Kiabi). 

Kiabi store is characteristic of an active ludic space. It is enclosed with a great number of micro- 

events (sub-area enclosed by high shelves, lots of shelves with different orientations). 

C&A store is a passive ludic space. Discontinuity is much lower than that of Kiabi. There are 

lots of micro-events but they introduce little discontinuity (shelves are not so long and not so 

high). 

Carrefour is a functional space. It is open on the outside and micro-events are not numerous and 

are predictable (straight, continued aisles with no breaks). 

156

Chart 3. Spatial configuration of the women’clothing area of Carrefour 

157

Glace 

Glace 

Chart 4. Spatial configuration of the women’clothing area of Kiabi 

158

Chart 5a. Spatial configuration of the women’clothing area of C&A (first floor) 

159

Chart 5b. Spatial configuration of the women’clothing area of C&A (second floor) 

To compare the number of appropriation strategies in each of these stores, we had to make sure 

that shopping contexts were the same. We, then, introduced certain control variables in our 

questionnaire (see chart 6): 

Shopping motives, either hedonic or utilitarian (Dawson and al, 1990) 

160

Shopping enjoyment (Schmidt and Spreng, 1996; Ohanian and Tashchian, 1992; Urbany and al, 

1996) 

Time pressure (Urbany and al, 1996; Putrevu and Ratchford, 1997) 

Planning of the purchase 

People accompanying the shopper 

Crowding (Hui and Bateson, 1991) 

Global shopping trip context (grocery shopping before or after the visit to the studied space, visit 

to another women’s clothing zone before or after the visit to the studied space). 

Spatial configuration 

Active ludic space 

Passive ludic space 

Functional space 

Appropriation strategies 

Active ludic strategies 

Passive ludic strategies 

Functional strategies 

Shopping contexts 

Shopping motives 

Shopping enjoyement 

Involvement 

Time pressure 

Planning of the purchase 

People accompanying the 

Chart 6. Frame of analysis for the study of spatial configuration’s influence on appropriation strategies 

Non participant and hidden observations were used. When the shopper went out of the zone, she 

was asked to answer the questionnaire. Data was collected during four weeks, seven half-days a 

week, Monday from 2 to 6 pm, Tuesday, Friday and Saturday from 10:30 am to 18:30 pm. 

726 usable observations and 450 questionnaires were collected. 251 concerned women shopping 

alone. 

There were 28 « going by » utilitarian strategies, 82 functional utilitarian strategies, 88 active 

ludic strategies and 53 passive ludic strategies. As mentioned, we will not consider the first 

strategy (“going by” functional). Only 223 observations were taken into account to test the 

hypotheses. 

161

No differences between the three different women’s clothing areas studied were found on the 

following control variables (see tables 6 and 7): shopping motives, shopping enjoyment, time 

pressure, involvement, crowding, grocery shopping trip before the visit to the studied space. 

For the remaining control variables, significant differences appeared. The results part of the 

paper therefore includes tests to check the influence of these variables on the results. 

Variables Utilitarian Hedonic Time pressure Crowding Involvement Involvement Shopping 

shopping shopping 

(interest and (sign) enjoyement 

motive motive 

pleasure) 

Signification ,224 ,360 ,381 ,468 ,089 ,226 ,532 

Table 6. Means comparison of the three women’s clothing areas for the control variables 

Variables Signification Contingency coefficient 

Grocery shopping before the visit 0.44 NV 

Grocery shopping after 0.000 0.28 

Visit to another women’clothing area before 0.008 0.2 

Visit to another women’s clothing area after 0.001 O.24 

Planning of the purchase 0.003 0.22 

Table 7. Comparison of the three women’s clothing areas. Khi-2 for the control variable (conditions fulfilled) 

II.2. Results 

Conditions of Khi-2 test are fulfilled for the comparison between spatial configuration and 

strategies. The differences are statistically significant (see table 8). The association is quite 

strong (contingence coefficient is 0.38). 

Spatial 

configuration 

Appropriation strategies Total 

Functional Active ludic Passive ludic 

Functional Number 43 11 21 75 

Theoretical 

number 

27,6 29,6 17,8 75,0 

Passive ludic Number 24 28 16 68 

Theoretical 

number 

25,0 26,8 16,2 68,0 

Active ludic Number 15 49 16 80 

Theoretical 

number 

29,4 31,6 19,0 80,0 

Total Number 82 88 53 223 

Theoretical 

number 

82,0 88,0 53,0 223,0 

Table 8.Khi-2 for appropriation strategies and spatial configuration (conditions fulfilled, Khi-2= 38.134 ; s =0.000 ; c = 0.382) 

- In the active ludic space, there are more active ludic strategies. H1 is accepted 

162

- In the passive ludic space, no particular strategy stood out. H2 is rejected 

- In the functional space, there are more functional utilitarian strategies. H3 is accepted. 

As we noticed differences in certain control variables, we checked if the results were still valid, whilst 

controlling these variables. We tested the hypothesis for each of the values of the control variables for 

which there was a difference between stores (table 9). 

Conditions of Khi-2 are not fulfilled in the case of planned purchase. In every other case, 

conditions are fulfilled. Significant differences are found for each value of every control 

variable. H1 and H3 are in all cases accepted. H2 is in all cases rejected. 

Variables Signification % theoretical number under 5 

Grocery shopping after 

Yes 0.01

appropriation strategies adopted by shoppers. On the contrary, the third spatial configuration 

(passive ludic space) does not seem to influence shoppers’ behaviour. 

Conclusion 

These two studies show that: 

-Four behavioural strategies can be distinguished, each of which is a means for the consumer to 

create a specific type of experience (“going by” functional, functional utilitarian, active ludic, 

passive ludic); 

-Discontinuity introduced in space creates the framework for shoppers’ mobility and influences 

the types of strategies shoppers choose and the experience created. High discontinuity 

encourages active ludic experience, whereas high continuity encourages functional utilitarian 

experience. 

As few empirical works were carried out on the topic, our two studies are exploratory. Two main 

limits can be mentioned. 

The first limit deals with the classification of the shopping trips. Few methodological tools exist 

to measure the reliability of the allocation of the shopping trips to the different strategies. 

Although we have used triangulation (with external data, with other research, and through 

several classifications), the comparison to the classification of the shopping trips by another 

researcher could be a useful indicator of reliability. Regarding the interpretation of the strategies, 

164

we used triangulation with other works on the same topic. Results would be further validated by 

using triangulation with shopper’s discourses about their own behaviour. 

The second limit concerns the use of women’s clothing zones of different retail chains. Some 

external factors (spaces surrounding the shopping area, types of products, atmospheric elements) 

may have challenged the validity of the study of spatial configuration’s influence on 

appropriation strategies. Although we have tried to neutralise the influence of these external 

factors by comparing the shopping contexts of the different spaces studied, confirmatory studies 

should compare the zones of stores of a same retail chain. 

The theoretical contribution of these studies is twofold. 

Firstly, they show that the physical environment-shopper interaction is not only mediated by the 

internal states of the person. Body mobility in the shopping environment is a key element in the 

creation of experience and attachment to a place. The study supports works that have proposed 

that consumers are active in the consumption process (Holt, 19995; Holbrook, 1999). Moreover, 

consumer research seems to focus on the study of internal states, neglecting physical behaviours’ 

analyse. Our studies illustrate the role of the body in the relational process with consumption 

objects (store being a specific consumption object). This work leads us to plead for a greater 

consideration of the role of the body in the understanding of the consumption process. 

Secondly, this research shows that spatial configuration is an influent dimension of physical 

environment. It influences shopping experience. Thus, the shopping environment does not only 

influence behaviour through sensorial stimuli. Store space management should then include two 

dimensions: one dealing with the management of sensorial cues, the other dealing with spatial 

configuration. 

165

A matrix of space management can then be proposed. Table 10 illustrates this with a few 

examples. 

Retailers should think about spatial configuration when they create store concepts. Using 

sensorial cues may not be efficient enough to create specific shopping experiences. The use of 

spatial configuration can reinforce the experience created and then the positioning of the store. 

The division of space in physical areas seem appropriate for “retailtainment” positioning. Grid 

aisles seem more appropriate for a functional positioning. 

Spatial 

configuration 

Active ludic 

(high discontinuity introduced 

by micro-events and enclosure) 

Functional 

(low continuity with little 

micro-events and openness) 

Management of sensorial cues 

Present Absent 

Sensorial active 

ludic space 

Nature & Découverte, 

Kiabi, Carrefour new 

concept, Nike Town, 

Ralph Lauren 

Sensorial 

functional space 

Cora (new concept) 

Non-sensorial 

active ludic space 

Troc de Lille, garage 

sales, flea markets 

Non-sensorial 

functional 

functional space 

Carrefour and Cora 

(early 1990’s concept) 

Table 10. A matrix of space management 

Two main research areas seem important to develop. 

Future research should endeavour to confirm the results of this work. Such studies should 

compare appropriation strategies in stores with different spatial configurations but belonging to 

the same retail chain. 

Other works could study the influence of the interaction between appropriation strategies, spatial 

configuration and shopping contexts on managerially relevant variables: consumption values, 

satisfaction, loyalty, buying behaviour. 

166

Bibliography 

Aubert-Gamet V. and Cova B. (1999), Servicescapes : From Modern Non-Places to Postmodern Common 

Places, Journal of Business Research, n°44, p.37-45 

Babin B.J. and Attaway J.S. (2000), Atmospheric Affect as a Tool for Creating Value and 

Gaining Share of Customer, Journal of Business Research, 49, 91-99 

Bitner M.-J. (1992), Servicescapes : The Impact of Physical Surroundings on Customers and Employees, 

Journal of Marketing, vol 56, April, 57-71. 

Bonnes M. and Secchiaroli G. (1995), Environmental psychology : a psycho-social introduction, Sage 

Publications, London 

Chebat J-C. and Robicheaux R. (2001), The interplay of emotions and cognitions of consumers 

in the retail environment, Journal of Business Research, 54, 87-88 

Chebat J-C., Chebat C.G. and Vaillant D. (2001), Environmetal background music and in store 

selling, Journal of Business Research, 54, 115-123 

Darden W.R. and Dorsch M.J. (1990), An Action Strategy Approach to Examining Shopping 

Behavior, Journal of Business Research, 21, p.289-308 

Dawson S., Bloch P.H. and Ridway N.M. (1990), Shopping motives, emotional states, and retail 

outcomes, Journal of Retailing, vol.66, n°4, Winter, p.408-427 

De Certeau M. (1980), L’invention du quotidien. Arts de faire, Gallimard, ed. 1990, Paris. 

Everett P. B., Pieters R. G. M. and Titus P. A. (1994), The consumer-environment interaction : An 

introduction to the special issue, International Journal of Research in Marketing, vol. 11, p. 97-105 

Fischer G.N. (1981), La psychosociologie de l’espace, Paris, PUF 

Fischer G.N. (1997), Psychologie de l’environnement social, Dunod, Paris. 

Floch J.M. (1990), Etes-vous arpenteur ou somnambule? L’élaboration d’une typologie comportementale 

des voyageurs dans le métro, in Sémiotique, marketing et communications, Floch J.M., Paris, PUF, p.19- 

48 

167

Holbrook M.B. (1999), Introduction to consumer value, in Holbrook M.B., Consumer Value, a 

framework for analysis and research, London, Routledge, interpretive marketing research series, p.1-28 

Holt D.B. (1995), How Consumers Consume : A Typology of Consumption Practices, Journal of 

Consumer Research, Vol. 22, Juin, p.1-16 

Hui M.K. and Bateson J.E.G. (1991), Perceived Control and the Effects of Crowding and Consumer 

Choice on the Service Experience, Journal of Consumer Research, Vol. 18, Septembre, p.174-184 

Kotler P. (1973), Atmospherics as a Marketing Tool, Journal of Retailing, vol. 49, n°4, p.48-64 

Kozinets R.V., Sherry J.F., DeBerry-Spence B., Duhachek A., Nuttavuthisit K. and Storm D. 

(2002), Themed flagship brand stores in the new millennium : theory, practice, prospects, 

Journal of Retailing, 78, p. 17-19 

Lam Y.S. (2001), The Effects of Store Environment on Shopping Behaviors : A Critical Review, 

Advances in Consumer Research, 28, 190-197 

Moles A. and Rohmer E. (1982), Labyrinthes du vécu. L’espace : matière d’actions, Librairie des 

méridiens, collection sociologies au quotidien 

Moles A.A. and Rohmer E. (1977), Théorie des actes. Vers une écologie des actions, Paris, Casterman 

Moles A.A. and Rohmer E. (1998), Psychosociologie de l’espace, textes rassemblés, mis en forme et 

présentés par Victor Schwach, L’Harmattan, Collection Villes et Entreprises, Paris 

Ohanian R. and Tashchian A. (1992), Consumers’ Shopping Effort and Evaluation of Store Image 

Attributes : The Roles of Purchasing Involvement and Recreational Shopping Interest, Journal of Applied 

Business Research, Vol.8, N°4, 40-49 

Putrevu S. and Ratchford B.T. (1997), A Model of Search Behavior with an Application to Grocery 

Shopping, Journal of Retailing, vol. 73, n°4, p.463-486 

Russel J.A. and Mehrabian A. (1976), Environmental variables in consumer research, Journal of 

Consumer Research, vol.3, June, p.62-63 

Schmidt J.B. and Spreng R.A. (1996), A proposed model of external consumer information search, 

Journal of the Academy of Marketing Science, vol.24, n°3, p.246-256 

168

Semprini A. (1990), Métro, Réseau, Ville : essai de sémiotique topologique, Nouveaux Actes 

Sémiotiques, 8, p.1-49 

Sherry Jr J.F. (1998), Introduction, in ServiceScapes : The Concept of Place in Contemporary Markets, 

Sherry Jr J.F, ed, NTC Business Books, Chicago, p.1-24 

Stokols D. (1978), Environmental Psychology, Annual Review of Psychology, 29, 253-259 

Titus P. A. and Everett P. B. (1995), The Consumer Retail Search Process : A Conceptual Model and 

Research Agenda, Journal of the Academy of Marketing Science, vol. 23, n°2, p. 106-119 

Turley L.W. and Chebat J-C. (2002), Linking Retail Strategy, Atmospheric Design and Shopping 

Behaviour, Journal of Marketing Management, 18, 125-144 

Turley L.W. and Milliman R.E. (2000), Atmospheric Effects on Shopping Behavior : A Review 

of the Experimental Evidence, Journal of Business Research, 49, 193-211 

Urbany J.E., Dickson P.R. and Kalapurakal R. (1996), Price search in the retail grocery market, Journal of 

Marketing, vol.60, April, p.91-104 

Véron E. and Levasseur M. (1991), Ethnographie de l’Exposition : l’Espace, le Corps et le Sens, BPI 

Centre George Pompidou, Paris 

169

Asymmetric Quality Tier Competition: An Alternative Explanation 

K. Sivakumar* 

ABSTRACT 

Research has provided extensive empirical support for the phenomenon of high quality brands benefiting 

more from price reductions than low quality brands. This article offers a new explanation for quality tier 

competition by using a specific mechanism for evaluating price-quality tradeoffs between brands in 

different tiers and demonstrates the results using scanner panel data. By offering an explanation using a 

simple analytical framework, it contributes to the literature by offering an alternative to the more involved 

explanations. 

170

ASYMMETRIC QUALITY TIER COMPETITION: AN ALTERNATIVE EXPLANATION 

Several reasons make a comprehensive understanding of the competition between brands of frequently 

purchased consumer products in different quality tiers academically interesting and managerially 

important. First, private label brands have been available for decades (Stern 1966) and their recent 

successes indicate that they are here to stay (Wall Street Journal 1993a, b). They contribute to more than 

20% of grocery sales (Marketing News 1995), and private label brands as a group have a higher market 

share than the top-selling national brands in 77 of 250 supermarket categories (Quelch and Harding 

1996). Second, retailers consider successful introduction of store label brands to be a natural extension of 

their power base (Wall Street Journal 1992). Third, the national brands affect and are affected by private 

label brands. Either their own plants manufacture for private labels or they introduce no-frills brands to 

compete with private label brands (Wall Street Journal 1993a). Finally, managers of many grocery stores 

(particularly chain stores) believe that a strong presence in the private label market can increase their 

profit as well as store loyalty of their customers (Richardson, Jain, and Dick 1996). 

Research has consistently shown asymmetry in quality tier competition. That is, by reducing prices, 

high quality brands are more successful in making consumers “switch up” from low quality brands than 

low quality brands are in making customers “switch down” from high quality brands. Several 

explanations have been offered for this phenomenon (e.g., Allenby and Rossi 1991; Blattberg and 

Wisniewski 1989; Hardie, Johnson, and Fader 1993). 

The objective of the research reported here is to offer an alternative explanation for asymmetric 

quality tier competition. The explanation, which is simpler than existing explanations, is based on 

consumer evaluation of product attribute tradeoffs. The article also offers empirical support using 

scanner panel data. 

SELECTIVE LITERATURE REVIEW 

Competition between two brands is asymmetric when one brand gains more from a price reduction than 

the other brand gains from an identical price reduction. A robust finding has been that high quality brands 

(e.g., national brands) benefit more than low quality brands (e.g., store label brands). Blattbergand 

171

Wisniewski (1989) proposed a theory of asymmetric quality tier competition based on consumer 

heterogeneity in reservation price differential. Allenby and Rossi’s (1991) explanation for asymmetric 

competition is based on an income effect and rotating indifference curves and the resulting nonhomothetic 

choice. They considered the high quality brand a superior good and the low quality brand an inferior 

good. Hardie, Johnson, and Fader (1993) argued that loss aversion for quality is more than loss aversion 

for price. Though asymmetric quality tier competition was not the focus of their study, they suggested 

that loss aversion implied asymmetric competition. 

Carpenter et al. (1988) used weekly market share data to estimate a regression model using zeta 

scores (Cooper and Nakanishi 1983, 1988). The higher priced brands offering some tangible benefits to 

consumers are the least affected by price competition from the economy brands. Higher priced brands that 

compete on the basis of image without having tangible benefits influence both the economy brand and the 

brands that offer tangible benefits. 

Using store level data, Kumar and Leone (1988) demonstrated that within a store, a high priced 

brand is able to influence the sale of a low priced brand more than vice versa. They further showed that 

high quality brands also have a stronger ability to induce store switching than low quality brands. 

Kamakura and Russell (1989) used disaggregate data at the household to show that some pattern of 

asymmetric preferences (people loyal to a given quality brand are willing to buy some quantity of higher 

quality brand but not lower quality) in the market are due to the presence of different segments of 

consumers, each of which behaves symmetrically toward all the brands in the market. Using store level 

weekly aggregated data, Allenby (1989) demonstrated a method to estimate the asymmetric competition 

among brands and then to identify a parsimonious structure of market competition by grouping similar 

products together and estimating a restricted model. One of the grouping schemes that Allenby (1989) 

considered was based on price tiers. His results show that high priced brands affect low priced brands 

more than vice versa. 

172

Sivakumar and Raj (1997) showed that asymmetry favors high quality brands in brand choice 

aspect as well as in category purchase aspect. They also showed that though high quality brands benefit 

more than low quality brands from a price decrease, the former are less vulnerable a price increase. 

CONCEPTUAL FRAMEWORK 

The objective of this section is to provide an intuitive rationale for the conceptual framework and to use 

the idea to derive a number of useful results. The explanation for asymmetric quality tier competition 

proposed here is based on product attribute tradeoffs. The basic argument is that while evaluating 

products, consumers consider tradeoffs between product attributes. That is, in the context of choosing 

brands from different quality tiers, a consumer considers how much extra quality (s)he gets for the 

differential price that (s)he pays. Support for the use of price/quality tradeoffs in our conceptual 

framework is provided from three different perspectives progressing from the general to the specific: (1) 

intuition, (2) support for the broad notion of tradeoff, (3) support for the specific notion of tradeoff. 

Any multi-attribute framework in consumer choice can be conceptualized in terms of tradeoffs 

among attributes (e.g., Guadagni and Little 1983; Blattberg and Wisniewski 1989). However, to formally 

incorporate tradeoffs between quality and price, we need to focus on the notion of differences in attributes 

gained per unit difference in price and consider it as a component in the consumer’s utility function. In 

the context of purchase decisions, comparing attribute differences between alternatives is intuitively quite 

appealing. Some example notions taken from everyday life are “by spending $150 extra, you can get an 

IBM computer with the same features rather than an Acer computer”; “spending just $50 more than a 

black and white printer, you can get a color printer”; and “by adding $20 to the basic membership fee, 

you will be able to add your spouse to the AAA membership.” In all these cases, in addition to the 

individual alternatives’ attributes, the decision is based on tradeoffs. Therefore, there is strong intuitive 

support for using tradeoff in determining consumer behavior. 

The formal notion of tradeoff as “what you get” divided by “what you pay” is a well known one. 

Researchers such as Monroe (1990) and Zeithaml (1988) and others have used it, albeit in the context of 

determining the attractiveness of individual alternatives first and then comparing the values across 

173

alternatives prior to the choice decision. The notion we use simply extends the principle for comparison 

of attribute differences across alternatives. More formally, there is evidence in the literature that people 

compare attributes of different alternatives in certain decision-making scenarios. The lexicographic 

decision rule involves comparing the attributes of different alternatives before deciding on the better 

alternatives (of course after determining that they tie on the other more important attributes). Examples of 

attribute comparison between alternatives also appear in other writings as well (e.g., Gale 1994). 

Therefore, from a conceptual perspective, comparing attributes between alternatives is reasonable. 

Finally, the specific rationale for considering tradeoffs in governing choice has been discussed in 

Simonson and Tversky (1992). Though their work focuses on attraction and compromise effects and 

changes in tradeoffs between background choice set and focal choice set (it does not consider changes in 

tradeoffs for a given choice set due to price changes nor are they concerned with price changes nor 

quality tier competition). The conceptual validity of our approach can be understood by their statement 

“... the choice between x and y then depends on whether the consumer is willing to pay $200 more for an 

additional 320K memory...” (Simonson and Tversky 1992 p. 282). Further, in their work on relative value 

theory (though unrelated to quality tier competition), Hollman and Lynch (1997) compare different 

formulations of attribute tradeoffs and find that the formulation similar to the one used in this article 

supports the choice process better than those based on other operationalizations. Thus, the notion of 

quality/price tradeoff has intuitive appeal, face validity, and literature support. 

The following discussion is based on Figure 1. For that framework, let us consider a two-brand 

market with brands denoted H and L. Let the price differential be denoted ‘b’ and the quality differential 

be denoted ‘a’ in the base condition of no price promotion. Note that quality means the quality perceived 

by the consumer. The proposed conceptual framework (consistent with other work in the literature) 

assumes that high quality brand is high priced and low quality brand is low priced). However, we do not 

assume that price is necessarily an indicator of quality in the short run (for example, when price reduction 

occurs, the quality is not assumed to change in the short run). Further, most work in this research stream 

assume the existence of price or quality tiers and also assume that high priced tier is of higher quality 

174

(that is in the larger, between-tier context). However, there is no assumption made regarding the price- 

quality association in the local or within tier context. This is consistent with available empirical evidence 

in the literature (e.g., Sivakumar and Raj 1997). 

__________________ 


__________________ 

In the base condition, the tradeoff in choosing H can be written a/b; that is, the consumer gets a 

quality advantage of ‘a’ for paying a price differential of ‘b’. Because the tradeoff is described in terms of 

quality differential per unit price differential, it is easy to see that increases in tradeoff help H and 

decreases in tradeoff hurt H (alternatively help L). For example, when the quality of H increase, the 

quality differential increases and hence, the numerator increases which benefits H. Note that because we 

are considering a two-brand market, H’s gain is L’s loss and vice versa. That is, whether we start with the 

premise that the tradeoff as expressed here helps H or hurts L, the conclusions will be identical. As the 

purpose is to demonstrate the intuitive value of the tradeoff concept, the discussion in this section is 

limited to two brands. The subsequent sections examine a number of multi-brand market configurations. 

When H reduces price by x, the price differential reduces to (b - x) and therefore the tradeoff 

increases to a/(b - x). In contrast, when L reduces price by x, the price differential increases to (b + x) and 

the tradeoff decreases to a/(b + x). Also note that the research stream on quality tier competition is 

predicated upon the existence of distinct tiers. Therefore, it is reasonable to assume that x 

is not, then the price of H will become less than the price of L. As the quality of H is more than that of L, 

the choice set is completely dominated by H and therefore not applicable in our context. Alternatively, if 

b is very small, there is no need to consider H and L as two different tiers but they can be considered 

similar to each other. Therefore, in all subsequent discussions, x is assumed to be less than b, which is 

reasonable for the context of our discussion. 

Based on the above discussion, the tradeoff in the base case is a/b, the tradeoff when H reduces 

price is a/(b - x), and the tradeoff when L reduces price is a/(b + x). Hence, the increase in tradeoff when 

H reduces price (in comparison with the base case) can be written as ∆H = a/(b - x) - a/b = ax/b(b - x). The 

175

decrease in tradeoff when L reduces price (in comparison with the base case) can be written as ∆L = a/b - 

a/(b + x) = ax/b(b + x). As (b + x) > (b - x), it is a mathematical fact that ∆H > ∆L. In other words, 

measured in terms of changes in tradeoffs, when H reduces price, the effect is stronger than that for a 

price reduction by L. Note that if we want to measure changes in tradeoffs relative to base tradeoff, both 

∆H and ∆L will only be adjusted by a scale factor represented by the tradeoff in the base scenario and 

therefore the conclusions derived subsequently will not change. 

Result #1: Price reductions benefit high-quality brands more than they do low-quality brands. 

Though we have used the generic word “benefit” in this section to demonstrate the intuitive nature 

of the results, subsequent sections demonstrate that these can be easily carried over to choice share 

scenarios. Note that Result #1 is consistent with overwhelming research evidence that asymmetry favors 

high quality brands (Allenby and Rossi 1991; Blattberg and Wisniewski 1989; Carpenter et al. 1988; 

Hardie, Fader, and Johnson 1993; Kamakura and Russell 1988; Sivakumar and Raj 1997). Consistency of 

that finding with previous research suggests that the proposed conceptual framework has face validity. 

A measure of the degree of asymmetric competition given by Z = ∆H - ∆L is a function of a, b, and 

x. After simplification, the degree of asymmetry can be denoted by 

(Equation 1) Z = 2ax 2 /b(b 2 - x 2 ). 

Because the asymmetry is defined in this way, it favors H if the measure is positive and it favors L 

if the measure is negative. Therefore increase in tradeoffs favors the choice of H and decrease favors L. 

Note that though our conceptual discussion focuses on “tradeoff” as the entity of interest, in an empirical 

sense, the tradeoff will (among other factors) affect the utility derived from the brands. Our conceptual 

development simply focuses on the differences in tradeoffs (after controlling for all other variables, as is 

often done in existing research) and therefore, the benefits as discussed in the conceptual development are 

differences in tradeoffs. This is just done for ease of exposition. As these tradeoffs are just one component 

of a brand’s utility, the weighting factor (i.e., the coefficient of the tradeoff variable) must be empirically 

determined (the utility will consist of a component given by the multiple of the tradeoff variable and the 

estimated coefficient). Incorporation of this coefficient and the resultant utility changes (rather than 

176

changes in tradeoffs as is done in the conceptual framework) will not affect the nature of the results 

because the weighting factor will only be a scale factor in the discussions and derivations. This is further 

demonstrated in the next section in which the discussion is based on choice theory framework. 

Though the theory development is based on the tradeoff variable itself, practical modeling of 

consumer choice often involves the modeling of utilities based on marketing variables and using the 

utility functions to determine the choice probabilities of various alternatives and the associated 

movements for price changes. The derivations and empirical analysis described subsequently incorporate 

the tradeoff variable as a part of the broader consumer utility function. 

Effect of Quality Differential on the Magnitude of Asymmetry 

A useful result that could be derived from the preceding expression for Z by differentiating it with respect 

to ‘a,’ the quality differential between the brands), is 

(Equation 2) dZ/da = 2x 2 /[b(b 2 -x 2 )] > 0. 

Result #2: Ceteris paribus, as the quality differential increases, the asymmetry favoring high 

quality brand increases. 

Products that give more value for the same price have a higher level of asymmetric advantage. As 

mentioned previously, ‘a’ can be visualized as a perceived quality differential. The perceived quality may 

be a function of many factors, including marketing management variables, consumer-specific variables, 

and product-specific variables. 

Effect of Price Differential on the Magnitude of Asymmetry 

Another result could be derived by differentiating Z with respect to the price differential. 

(Equation 3) dZ/db = -2ax 2 (3b 2 -x 2 )/(b 3 -bx 2 ) 2 < 0. 

Result #3: Ceteris paribus, as the price differential increases, the asymmetry favoring high quality 

brand decreases. 

Therefore, for a given quality differential, increasing the price of a high quality brand will be 

detrimental to the favorable asymmetry toward it. Here again, the price differential can be considered a 

perceived price differential. 

177

As Bronnenberg and Wathieu (1996) use loss aversion (first proposed by Hardie, Johnson, and 

Fader 1993) to explain asymmetric competition as opposed to the explanation based on price/quality 

tradeoffs in our paper, the two works are not directly comparable. This is because from a conceptual 

perspective, while tradeoff is a fundamental process that gives rise to asymmetry in our paper, the 

expression happens to be the result arising from the loss aversion phenomenon in Bronnenberg and 

Wathieu (1996). Proposition 1 derived in their paper, though more detailed, basically implies that as the 

quality differential goes up, the asymmetry favoring H (L) increases (decreases) and as the price 

differential goes up, the asymmetry favoring H (L) decreases (increases). This conclusion is consistent 

with the results derived in our paper (Result #2 and Result #3) which also state that increase in quality 

differential enhances asymmetry favoring H and increase in price differential subdues the asymmetry 

favoring H. Therefore, though the explanations and the specificity of results for asymmetric competition 

are different between our paper and Bronnenberg and Wathieu’s (1996) work, there is consistency in the 

outcome of the two explanations, thereby providing further face validity for our framework. Interestingly, 

we are able to derive qualitatively similar results without resort to a strong assumption of loss aversion. 

Further, as described below, our framework also enables the derivation of some useful results related to 

pricing strategies for different quality tiers. This is also a new feature that distinguishes our research from 

other research on quality tier competition. 

A key contribution of the conceptual framework is that it offers a new explanation for the main 

effect prediction of asymmetric quality tier competition that is consistent with research in this area 

(Allenby and Rossi 1991; Blattberg and Wisniewski 1989; Hardie, Johnson, and Fader 1993; Sivakumar 

and Raj 1997). A positive feature of this new explanation is that the hypotheses are based on the well- 

accepted notion of tradeoffs and hence based on weaker (i.e., less restrictive) assumptions compared to 

stronger (i.e., more restrictive) assumptions used in existing research. For example, Blattberg and 

Wisniewski’s (1989) explanation is based on a particular distribution of reservation prices, Allenby and 

Rossi’s (1991) explanation is based on rotating indifference curves, and Hardie, Johnson, and Fader’s 

(1993) explanation is based on loss aversion – all of which are stronger and more restrictive assumptions 

178

than the one based on the notion of tradeoff used in this research. Furthermore, the two additional results 

(Result #2 and Result #3), though not exactly the same, are consistent with the findings of Bronnenberg 

and Wathieu (1996). 

EMPIRICAL ILLUSTRATION 

The key contribution of this manuscript is in the simplicity of the alternative explanation, and the 

comprehensive way in which the intuitive notion of tradeoff appears to apply in a wide variety of market 

configurations. The key aspect of any potential empirical analysis is that all the results follow directly 

from the positive coefficient for the tradeoff variable (therein also lies the simplicity and contribution of 

this manuscript). Thus, empirical analysis in this article is somewhat different compared to other articles. 

However, the exact quantitative impact of the various price changes must be verified using real numbers. 

Given the nature of the results derived and the notion that it is not possible to obtain all possible scenarios 

and all possible operationalizations of tradeoff variables using one data set, the objectives of the empirical 

analysis are rather specific and threefold. The first objective is to show that the findings are valid in real- 

life situations. The second objective is to get an idea about the magnitudes of the various coefficients so 

that the exact impact of price changes can be quantified. 

Model 

The well-known logit model of brand choice (Ben-Akiva and Lerman 1985; Guadagni and Little 1983; 

McFadden 1981) is used for empirical illustration of the proposed conceptual framework. In the logit 

model, the choice probability of each alternative is based on the utility derived from that alternative. If Ui 

denotes the utility derived from brand i, then the probability of choosing the alternative i can be written as 

e Ui /Σj e Uj . The method is developed by using a two-brand market consisting of one high quality brand 

H and one low quality brand L (without loss of generality), though the empirical analysis incorporates 

multiple brands in the market. The utilities for brands H and L can now be written as 

UH = QH + VH + βPH + τ(QH - QL)/(PH - PL) + εH; UL = QL + VL + βPL + εL. 

Where QH, QL = quality of H and L, respectively; VH, VH = brand specific utilities not accounted by other 

variables included in the model; PH, PL = prices of H and L respectively. 

179

Those conceptually based utility functions must be reduced to estimable form for the estimation of 

appropriate coefficients in the model. Further, some of the coefficients must be normalized to zero to 

make the model estimable. 

The above utility functions can be rewritten as UH = αH + βPH + γ/(PH - PL) and UL = αL + βPL, 

where αH = QH + VH and γ = τ(QH - QL) and αL = QL + VL. 

For logit model, a constant terms must be normalized to zero. Therefore, the components of the 

model to be estimated (after exclusion of the variables other than those involving price) will be 

UH = αH + βPH + γ/(PH-PL); UL = βPL. 

Note that the exact determination of QH - QL is not necessary to understand the combined effect of 

price including the tradeoff variable. The reason is that to assess the effect of price, all we need are the 

coefficients associated with variables containing price in one form or another. 

Data 

The dataset used in the empirical analysis is purchase data provided by IRI under the auspices of the 

Marketing Science Institute. In addition to the faced price, manufacturers’ coupon and store coupon 

redemption were recorded, as well as in-store features and displays and competing brand prices. A 

demographic file contains relevant information on family size, income, and so on. The purchase records 

covered a two-year period for a large retail chain. On the basis of the first 26 weeks, families not 

purchasing at least once in two months were excluded. The four top brands are included in the analysis, 

three national brands (denoted H1, H2, and H3) and one store label brand (denoted L). H2 and H3 are minor 

brands. In addition to the four-brand analysis reported here, an analysis was conducted without the two 

minor brands. The conclusions are fully consistent with the conclusions drawn here. Therefore, to 

conserve space, only the results of the four-brand analysis are reported and discussed. The data 

management procedures are similar to those followed in existing research using similar data sets. Data for 

purchase occasions during the first one third of the time period were used to initialize the brand loyalty 

variable (described subsequently) and the remaining data were used to estimate the models. 

Model Variables 

180

The variables incorporated in the model were based on previous research on choice models using scanner 

panel data. However, an important addition is a variable to incorporate the tradeoffs. Demographic 

variables are included to account for some heterogeneity among consumers (we estimate models with and 

without demographic variables and the results are consistent). The proposed formulation is compared with 

models not incorporating the tradeoffs by means of likelihood ratio tests. Managerial implications can be 

delineated by calculating the inter-brand movements after incorporating the tradeoff construct. The 

deterministic components of the utilities for several models follow. 

Model 1 (model with price, but without tradeoff) 

UHi = αHi + θ1DisplayHi + θ2FeatureHi + θ3iFamilysize + θ4iIncome + θ5BrandloyaltyHi + βPriceHi 

UL = θ1DisplayL + θ2FeatureL + θ5BrandloyaltyL + βPriceL. 

Model 2 (model without price, but with tradeoff) 

UHi=αHi+θ1DisplayHi+θ2FeatureHi+θ3iFamilysize+θ4iIncome+θ5BrandloyaltyHi+γ/(PriceHi-PriceL) 

UL = θ1DisplayL + θ2FeatureL + θ5BrandloyaltyL. 

Model 3 (model with price and tradeoff) 

UHi=αHi+θ1DisplayHi+θ2FeatureHi+θ3iFamilysize+θ4iIncome+θ5BrandloyaltyHi+βPriceHi 

+γ/(PriceHi- PriceL) 


Model 4 (parsimonious model with price and tradeoff) 

UHi = αHi + θ1DisplayHi + θ2FeatureHi + θ5BrandloyaltyHi + βPriceHi + γ/(PriceHi - PriceL) 


Operationalization of Variables 

Model variables were patterned after existing research investigating substantive topics on quality tier 

competition. The idea is to use logically consistent operationalizations that have been utilized in existing 

research. Display and feature are categoric variables coded as 1 if present and 0 if absent. Price is the 

faced price (for the non-purchased brands) or paid price (for the purchased brands) as has been used in 

181

ecent existing research (Sivakumar and Raj 1997). Brand loyalty is defined as the cumulative proportion 

of purchases devoted to the brand until the current purchase occasion (Sivakumar and Raj 1997). Family 

size and income are ordinal variables that capture family income and size. Finally, the price differential 

(for the tradeoff variable) is the difference in actual paid prices. 

RESULTS 

The estimation results for the models are reported in Table 1. The coefficients are of the appropriate sign 

and the relevant variables contribute significantly toward explaining brand choice. Model fit can be 

assessed by comparing the loglikelihoods or U 2 . Nested models could be compared by using a likelihood 

ratio test. In nested models, twice the difference in loglikelihood between two models is distributed as chi 

squared with degrees of freedom given by the difference the numbers of model parameters. 

___________________ 


___________________ 

The model incorporating the tradeoff variable did significantly better than the model not 

incorporating it. Therefore, the tradeoff variable improves the fit considerably. It is interesting to note that 

after incorporation of the tradeoff variable, the inclusion of price does not improve model fit significantly. 

Hereafter, we will only discuss model M4. These computations are provided in Table 2. 

___________________ 


___________________ 

A measure of asymmetry is needed for the proposed logit choice model to test the analytical results 

derived earlier. Mela, Gupta, and Lehmann (1997) and Sivakumar and Raj (1997) respectively note that 

cross elasticity may not be an appropriate measure for depicting price effects in general and asymmetry in 

particular. They note that a measure of interbrand movement or its variant is needed to accurately 

understand asymmetry (i.e., asymmetry favors H if the movement from L to H for a given price reduction 

by H is more than the movement from H to L for the same price reduction by L). Therefore, to measure 

the extent of asymmetry, we need to compare movements for equal price changes by H and L 

respectively. Based on the general approach suggested by Mela, Gupta, and Lehmann (1997) and 

182

Sivakumar and Raj (1997), inter-brand movements can be derived analytically as we demonstrated earlier 

in the theory development. However, analytical derivations are based on infinitesimal calculus. While 

useful to theorize, in order to get an idea of actual movements and also to see whether it works with real 

numbers, we need to construct realistic scenarios. Therefore, share changes are computed by 

incorporating various values in the utility functions and making suitable assumptions to ensure that the 

market share and other key market characteristics are kept close to reality as possible. These computations 

presented in Table 2 demonstrates that between any pair of high-low tier brands (i.e., H1 and L, H2 and L, 

and H3 and L), the results derived earlier holds good. Note that the asymmetry holds good for all 

combinations irrespective of their market share. Also, to avoid overcrowding of the table, some of the 

intermediate calculations discussed below are not given in the table but the final results are. 

Evidence for Asymmetric Quality Tier Competition. Result 1 derived from the conceptual framework 

implies that the competition between a high quality brand and a low quality brand is asymmetric, 

consistent with the predominant research findings in marketing. For example, Table 2 indicates a 

movement of 6.26 market share points from L to H1 when H1 reduces its price by 20 cents and a 

movement of 3.78 market share points from H1 to L when L reduces its price by 20 cents. Clearly 

asymmetry favors H1 and the asymmetric advantage amounts to 2.48 market share points. The asymmetric 

advantage of H2 over L is 0.53 and that of H3 over L is 0.70. 

Role of Price Differential. Ceteris paribus (i.e., for a given quality differential), increasing the price 

differential was hypothesized to result in smaller asymmetric advantage for high quality brands (H2). 

Table 2 illustrates the asymmetry for the base scenario and the scenario in which the price differential is 

increased 10 cents. The larger price differential results in an asymmetry (favoring H1 over L) of 1.58 

market share point versus asymmetry of 2.48 market share points in the base case (a difference of 0.9), 

confirming result 2. Similar results hold for competition between H2 and L and that between H3 and L. 

Role of Quality Differential. Ceteris paribus (i.e., for a given price differential), increasing the quality 

differential was hypothesized to result in greater asymmetric advantage for high quality brands (Result 3). 

Table 2 illustrates the asymmetry for the base scenario and the scenario in which the quality differential is 

183

increased by 10% (conceptually the same as increasing the value of γ by 10%). Consistent with result 3, 

the larger quality differential results in asymmetry (favoring H1 over L) of 2.74 market share points 

versus the asymmetry of 2.48 market share points in the base case (a difference of 0.26). 

DISCUSSION 

Managerial Implications. An important implication of the proposed framework is that arbitrarily 

increasing the price of high quality brands without a corresponding increase in their perceived quality 

may not be a desirable strategy. Price increases should be accompanied by perceived improvements in 

quality. This conclusion is consistent with Bronnenberg and Wathieu (1996) who use a different 

conceptual framework to explain asymmetric competition. A second implication is that the tradeoff 

between frequency and depth of price cut is not uniform across all brands, but is contingent upon the 

perceived quality level of the brands. 

The primary outcome suggested by all existing frameworks is that quality tier competition is 

asymmetric (though the degree of asymmetry is a function of various factors). However, the rationale for 

such asymmetry has implications for how a company tries to achieve superior performance in the 

marketplace. Blattberg and Wisniewski’s (1989) explanation is based on consumer heterogeneity. 

Therefore, the asymmetry does not depend on individual customers but on a particular pattern of 

heterogeneity. The explanations by Allenby and Rossi (1991), Hardie, Johnson, and Fader (1993) and 

Bronnenberg and Wathieu (1996), as well as the research reported here, are based on individual-level 

phenomena. 

Though most frameworks in the literature have only focused on the nature of quality tier 

competition (Result #1), the other results derived here are consistent with the conceptual rationale of the 

other explanations for asymmetric competition. 

Research Implications. Much work needs to be done in terms of formal comparison of the four 

frameworks, as well as to understand the psychological processes involved in decision making when 

consumers chose between brands. An interesting research issue is the way consumers (seem to) 

operationalize the tradeoff variable. Though a positive feature of this work is the demonstration of the 

184

esults for a number of operationalizations, it would be useful to identify segments of consumers who use 

different methods so that a more detailed understanding of the market place is obtained. Research based 

on in-depth interviews and other qualitative methods such as process tracing may shed more light on the 

processes leading to choice. Also, different conceptual frameworks may be appropriate for different 

product categories or different sets of consumers. Investigations of those issues should lead to a more 

detailed understanding of the competition between different quality tiers. 

Interestingly, all research on asymmetric quality tier competition in a brand choice context is based 

on scanner panel data rather than the direct measurement of consumer perceptions. Also, studies on 

consumer perceptions of store label brands (e.g., Bettman 1974; Burger and Schott; Coe 1991; Frank and 

Boyd 1965; Richardson, Jain, and Dick 1996) do not model the competition between quality tiers. Hence, 

the research domain could be meaningfully extended by a triangulation of those two research streams. 

Another useful avenue for future research is meta-analytical integration of research on 

quality tier competition. Sethuraman (1995) attempted a statistical summary of 261 cross-price 

elasticities estimated in his article. However, meta-analysis is needed on the evidence from other 

studies demonstrating asymmetry in quality tier competition. Similarly, more work is needed in 

understanding and delineating the boundary conditions of asymmetric quality tier competition. 

185

REFERENCES 

Allenby GM. A unified approach to identifying, estimating and testing demand structures with aggregate 

scanner data. Mark Sci 1989;8(3):265-280. 

Allenby GM, Rossi PE. Quality perceptions and asymmetric switching between brands. Mark Sci 

1991;10(3):185-204. 

Ben-Akiva M, Lerman S. Discrete choice analysis. Cambridge (MA): M.I.T. Press, 1985. 

Bettman JR. Relationship of information-processing attitude structures to private brand purchasing 

behavior. J Appl Psychol 1974;59(1):79-83. 

Blattberg RC, Wisniewski KJ. Price-induced patterns of competition. Mark Sci 1989;8(4):291-310. 

Bronnenberg B, Wathieu L. Asymmetric promotion effects and brand positioning. Mark Sci 

1996;15(4):379-394. 

Burger PC, Schott B. Can private brand buyers be identified? J Mark Res 1972;9(2):219-222 (May). 

Carpenter GS, Cooper LG, Hanssens DM, Midgley DF. Modeling asymmetric competition. Mark Sci 

1988;7(4):393-412. 

Coe B. Private versus national preference among lower and middle-income consumers. J Retailing 

1971;4(3):61-72. 

Fader PS, Lattin JM. Accounting for heterogeneity and nonstationarity in a cross-sectional model of 

consumer purchase behavior. Mark Sci 1991;12(3):304-317. 

Gale BT. Managing customer value. New York (NY): The Free Press, 1994. 

Guadagni PM, Little JDC. A logit model of brand choice calibrated on scanner data. Mark Sci 

1983;2:203-38 (Summer). 

Hardie B, Johnson E, Fader P. Modeling loss aversion and reference dependence effects on brand choice. 

Mark Sci 1993;12(4):378-394. 

Hoch SJ. How should national brands think about private labels? Sloan Manage Rev 1996;37(2):89-96. 

Hoch SJ, Dreze X, Purk M. EDLP, hi-lo, and margin arithmetic. J Mark 1994;58(4):16-29 (October). 

Hollman F, Lynch J. Relative value theory: an analytic model of contextual choice analysis. Working 

Paper, University of Florida, 1997. 

Kamakura WA, Russell GJ. A probabilistic choice model for market segmentation and elasticity structure. 

J Mark Res 1989;26(4):379-390 (November). 

Kumar V, Leone RP. Measuring the effect of retail store promotions on brand and store substitution. J 

Mark Res 1988;35(2):178-185 (May). 

Lal R, Rao R. Supermarket competition: the case of every day low pricing. Mark Sci 1997;16(1):60-80. 

186

Mela CF, Gupta S, Lehmann DR. Long term impact of promotion and advertising on consumer brand 

choice. J Mark Res 1997;35(2):248-261 (May). 

Monroe KB. Pricing: making profitable decisions, 2nd edition. New York (NY): McGraw-Hill Publishing 

Company, 1990. 

Quelch JA, Harding D. Brands versus private labels: fighting to win. Harvard Bus Rev 1996;74(1):99-109 

(January-February). 

Raju JS, Srinivasan V, Lal R. The effects of brand loyalty on competitive price promotional strategies. 

Manage Sci 1990;36(3):276-304 (March). 

Richardson PS, Jain AK, Dick A. Household store brand proneness: a framework. J Retailing 

1996;72(2):159-185. 

Sethuraman R. A meta-analysis of national brand and store brand cross-promotional price elasticities. 

Mark Lett 1995;6(4):275-286. 

Sethuraman R. A model of how discounting high-priced brands affects the sales of low-priced brands. J 

Mark Res 1996;33(4):399-409 (November). 

Simonson I, Tversky A. Choice in context: tradeoff contrast and extremeness aversion. J Mark Res 

1992;29(3):281-95 (August). 

Sivakumar K, Raj SP. Quality tier competition: how price change influences brand choice and 

category choice. J Mark 1997;61(3):71-84 (July). 

Stern L. The new world of private labels. California Manage Rev 1966;8(3):43-50. 

Wall Street Journal, The. Soft drink giants sit up and take notice as sales of store brands show more fizz. 

1992, March 6, B1. 

Wall Street Journal, The. Brand names are getting steamed up to peel off their private-label rivals. 1993, 

April 21, B1 and B5. 

Wall Street Journal, The. Advertisers seek to save brands from price cuts. 1993, April 21, B1 - B6. 

Zeithaml VA. Consumers’ perceptions of price, quality, and value: a means-end model and synthesis of 

evidence. J Mark 1988;52(3):2-22 (July). 

187

Table 1. Estimation Results 

____________________________________________________________________________________ 

Variables Model M1 Model M2 Model M3 Model M4 

(Price only) (Tradeoff only) (Price & Tradeoff) (Reduced / 

Price & 

Tradeoff) 

____________________________________________________________________________________ 

Brand constant (A) 0.896* -1.067** -0.716 -0.683 

Brand constant (B) 1.353** -0.483 -0.145 -0.579 

Brand constant (C) 0.575 -1.884*** -1.509** -2.107*** 

Display 1.685*** 1.597*** 1.564*** 1.557*** 

Feature 1.240*** 1.006*** 0.932** 0.904** 

Family size (A) -0.036 -0.032 -0.030 

Family size (B) 0.015 0.029 0.352 

Family size (C) -0.034 -0.042 -0.038 

Income (A) 0.022 0.034 0.038 

Income (B) -0.151 -0.142 -0.141 

Income (C) -0.132 -0.111 -0.111 

Brand Loyalty 4.694*** 4.654*** 4.675*** 4.707*** 

Price -0.0195*** -0.0054 -0.0051 

Tradeoff 31.117*** 28.52*** 28.836*** 

____________________________________________________________________________________ 

Model χ 2 (d.f.) 2330.9 (13) 2353.0 (13) 2353.6 (14) 2348.7 (8) 

Model Log-likelihood -297.1 -286.07 -285.74 -288.29 

Model Fit (U 2 ) 0.797 0.804 0.805 0.803 

____________________________________________________________________________________ 

* significant at p < 0.10; ** significant at p < 0.05; *** significant at p < 0.01. 

Relevant model comparisons based on likelihood ratio tests: 

(a) Incorporation of Tradeoff helps to improve fit (M3 better than M1 at p < 0.01). 

(b) Reduced models are not significantly worse than the full models (M4 not worse than M3 at p > 0.1). 

(c) After incorporating the tradeoff term, the straight price does not improve the fit of the model 

significantly (M3 not better than M2). 

188

Table 2. Empirical Illustration of Aspects of Quality Tier Competition 

______________________________________________________________________________ 

____________________________ 

Modifications to 

the Base Scenario 

Tradeoff Price 

Quality Base Scenario 

Coefficient Differential Differ. 

Increases Increases 

Increases 

______________________________________________________________________________ 

____________________________ 

Competition between H1 and L 

Degree of asymmetry (H1 over L) 

3.00 

1.58 

2.74 

Competition between H 2 and L 

Degree of asymmetry (H 2 over L) 

0.65 

0.35 

0.59 

Competition between H 3 and L 

Degree of asymmetry (H 3 over L) 0.70 0.87 

0.43 0.78 

____________________________________________________________________________________________ 

_____ 

Figure 1. Attribute Tradeoffs as an Explanation for Quality Tier Competition 

189 

2.48 

0.53

Q 

U 

A 

L 

I 

T 

a 

H H’ 

x 

b 

PRICE 

190 

L L’ 

x

Modeling The Impact Of Internet Atmospherics On Surfer Behavior 

Marie-Odile Richard, École des Hautes Études Commerciales 

Abstract 

The goal of this study was to examine the impact of atmospherics cues concerning Internet 

advertising on the behavior of consumers and its impact on several variables such as attitudes, 

involvement, exploratory behavior, pre-purchase and purchase intentions. Atmospherics cues 

were central (structure, informativeness, effectiveness and navigational characteristics) and 

peripheral (entertainment). A conceptual model is developed based on a review of existing 

findings and tested with a random sample of real consumers who responded to a questionnaire 

after navigating through an existing pharmaceutical web site. Structural equations modeling was 

used to test nine major hypotheses. Among key findings, all atmospherics cues were impacting 

the other constructs, with the central cues mostly affecting involvement and exploratory 

behavior, while entertainment affected involvement and attitude. These findings contribute to 

theoretical and managerial understanding of the role of Internet atmospherics on the web 

navigation behavior of visitors. 

191

MODELING THE IMPACT OF INTERNET ATMOSPHERICS ON SURFER BEHAVIOR 

1. Introduction 

The Internet is growing in importance since the concept of a product is becoming more 

information-based (Brännback, 1997) and offers the opportunity to separate information about a 

product (or service) from the product itself (Rayport and Sviokla, 1994). 

A 1997 study of URLs in mass media ads showed that nearly one-fourth of the non-users of 

the Internet were planning to log on within the next month, and their principal destination was a 

health site (Maddox et al., 1997; Maddox, 1999). For current users of the Internet, health-related 

sites ranked second as a destination within the next month, with 51% planning to visit one site 

(Maddox et al., 1997). Of those who seek medical data online, 52% are looking for information 

on diseases and 33% on pharmaceuticals (Miller and Reents, 1998). As Internet usage continues 

to grow across the world, the focus for many companies is slowly shifting from merely creating 

to more strategic aspects involving how to best use this medium. One of the most important 

objectives of firms on the Web remains effective communications with consumers. This 

emphasizes the importance of developing and testing systematic models of the Web as an 

advertising or communication tool. 

Most research on online communication of information is in the context of online retailing. This 

study seeks to expand the scope of models of consumer behavior and responses to Web site and interface 

characteristics. More explicitely, it proposes and empirically tests a general model of consumer behavior 

and responses to Web site environmental cues based on theoretical frameworks for Web atmospherics in a 

retail setting developed by Mehrabian and Russell (1974); Donovan and Rossiter (1982) and, more 

recently, Dailey (2002) and Eroglu, Machleit and Davis (2001). 

Our study differs from and extends the literature in several ways: first, unlike most previous research, 

we propose and test our model in the general context of information-seeking in online retailing 

specifically. Second, in the spirit of Eroglu et al. (2001) we provide a broader, more comprehensive 

model integrating research on Web site environmental cues (to stimulate the search for information), 

internal states conceptualized by cognitive and affective aspects of the surfers’ behavior, and finally 

outcomes. Third, our study goes beyond theoretical model development; our empirical testing is 

performed in the context of a real-world pharmaceutical Web site. Finally, we apply a significantly more 

sophisticated methodology that allows us to test the structure and validity of our proposed model. 

We begin by briefly discussing previous literature relevant to the antecedents, processing and 

consequences of interest to our model. This is followed by the presentation of our proposed 

192

model along with development of the hypotheses. The following section describes the empirical 

approach and the data used to test the model. We then present the results of empirical testing, 

and conclude with a discussion of the finding, implications and limitations. 

2. Conceptual Background 

Turley and Milliman (2000) identified over 60 published studies with significant 

relationships between store atmospherics and consumer behavior. Atmospherics has been shown 

to influence consumer perceptions of retail products (Obermiller and Bitner, 1984), and store 

approach/avoidance behaviors such as consumers' intention and decision to frequent and spend 

money in a store (Turley and Milliman, 2000; Donovan and Rossiter, 1982; Donovan et al., 

1994; Darden et al, 1983; Stanley and Sewall, 1976). 

Kotler (1973) defined atmospherics as “the intentional control and structuring of environmental cues” 

or as “the conscious design of space to create certain buyer effects.” Milliman and Fugate (1993) defined 

atmospheric cues as “any component within an individual’s perceptual field that stimulates one’s senses.” 

Atmospheric cues may be more influential than other marketing inputs that are not present at the point of 

purchase (Baker, 1986; Baker et al., 1994) and be more influential in the purchase decision than the 

product itself (Kotler, 1973). 

To explain the influence of atmospherics on consumers, the Mehrabian–Russell Affect 

Model (Mehrabian and Russell, 1974; Turley and Milliman, 2000), is based on Stimulus- 

Organism-Response paradigm. Features of the environment (S) cause behavioral responses 

(liking the store, enjoyment of shopping in it, willingness to spend time in it, to explore the 

environment, or to return to it, likelihood of spending more money than intended, feelings of 

friendliness to others) (Donovan et al, 1994) within it (internet, in our case) by altering consumer 

affect emotional states (O) (such as pleasure, arousal and sometimes dominance) (Mehrabian and 

Russell, 1974; Donovan and Rossiter, 1982), depending on individual personality traits analyzed 

in the Zuckerman‘s sensation-seeking scale (Grossbart et al., 1975). 

By changing Kotler’s (1973) definition of brick-and-mortar atmospherics, Web atmospherics 

can be defined as the “conscious designing of Web environments to create positive effects 

(positive affect and/or cognitions) in surfers in order to develop positive consumer responses 

(site revisiting, browsing)” (Dailey, 2002). For Milliman and Fugate (1993), a Web atmospheric 

cue is comparable to a brick-and-mortar atmospheric cue and can be described as “any Web 

interface component within an individual's perceptual field that stimulates one's senses.” 

193

Eroglu et al (2001) proposed a typology that divides Web atmospheric tools in two groups: 

high task-relevant and low task-relevant cues. On one hand, high task-relevant cues facilitate and 

enable the consumer's shopping goal attainment (i.e., descriptions or pictures of the merchandise, 

price, navigation cues, etc.). On the other hand, low task-relevant cues are defined as being 

inconsequential to the completion of the shopping task (i.e., colors, borders and background 

patterns, typestyles and fonts, animation, music and sounds, entertainment, decorative pictures, 

etc.) (Eroglu et al., 2001). 

Unfortunately, Web atmospheric research has been limited to why atmospherics influence 

surfers. In fact, researchers have not examined specific atmospheric cues (i.e., colors, layout, 

navigation cues, etc.), but have focused on general atmospheric cues (high task-relevant cues, 

positive atmospheres), decreasing the probability of finding theories that adequately explain the 

unique influence of these cues (Turley and Milliman, 2000). 

Donovan and Rossiter’s (1982) study used students subjects and measured attitudes and 

intentions rather than shopping behavior. They did not assess the contribution of emotional 

factors relative to cognitive factors except by Anderson (1986). The Donovan, Rossiter and 

Marcoolyn’ study (1994) further looked at the extent to which the Mehrabian-Russell model 

predicted the cognitive factors (e.g., perception of merchandise quality, variety, value for 

money) independently of affective factors. It is important to analyze emotional states 

independently from cognitive variables and both together. 

Two types of involvement influence the motivation to attend and process product information. In our 

study, we assumed that situational involvement is present as our sample is constituted of students. 

Situational involvement for the product is more likely to result in a goal-directed behavior (temporary, 

considered linked to a short-term visit and occurs only in the context of a situation). 

The Elaboration Likelihood Model (ELM) suggests that low-involvement subjects process 

information through the peripheral rather than the central route (Petty, Cacioppo and Schumann, 1983b), 

relying more heavily on cues as opposed to detailed and elaborate product specific information. With the 

Internet, marketers can use a large number of cues (e.g., search engines, keywords, search arguments) to 

attract and influence consumers (McGaughey and Mason, 1998). 

In conclusion, in the Web medium, consumers choose the amount of exposure to a Web site, 

and decide what to watch, when, and how much. If they find that the Web site is not useful or 

interesting, they will immediately end their visit. Attitude is positive when there is ease of 

navigation through a database or ability to get an overview of the structure of the site. It leads to 

an increase of the number of visits to the site, in curiosity, in depth of search, in exploratory 

194

ehavior and in duration time spent. Source attractiveness also influences the amount of 

information processing. Effectiveness of information content develops emotional responses such 

as pleasure and arousal that in turn mediate a variety of approach-avoidance behaviors. The latter 

are characterized as a general liking of the environment; attitude toward returning to the same 

Web sites; shopping time and in-depth exploration of the Web sites. 

3. Conceptual Model 

Following MacInnis and Jaworski (1989), our model, represented in figure 1, can be divided into 

three parts: antecedents, processing, and consequences. More specifically, as reported by Mehrabian and 

Russell (1974) and Donovan and Rossiter (1982), it can also be divided into three parts: Stimuli, 

Organism and Outcomes. Dailey (2002) mentioned two types of stimuli: high relevant cues and low- 

relevant cues. We chose to place most of our environmental variables related to the Internet (i.e., 

navigational cues, structure and informativeness of the site, its entertainment quality and effectiveness of 

its information content) in the first category because our concern is to evaluate the impact of the 

information content on the other variables. In the second category, figured entertainment of the Web site, 

which is linked to the first category, but indirectly. These dimensions lead to the processing variables 

such as approach/avoidance behaviors (attitudes), an affective variable, and exploratory behavior and 

involvement categorized as cognitive variables (i.e., related to information search process). We place the 

focus on affective and cognitive variables that have been proposed for the model in figure 1 to test if it is 

applicable to the Internet. We complete this model by the outcomes such as pre-purchase intentions (or 

involvement in purchase decisions) and purchase behavior related to the product. 

[Insert figure 1 around here] 

We now describe our model in more detail, and systematically develop all the hypotheses that will be 

tested in the study. 

3.1 Internet Atmospherics (Environmental) Cues 

Internet atmospherics cues, the central focus of our model, are critical to the effectiveness of a web 

site since it is entirely in the customers’ power to decide which Web pages to browse, for how long, and 

how much information to obtain. 

In an attempt to develop a better understanding of what constitutes high-quality Web content, we 

propose several factors that are important for surfers to evaluate when navigating through a Web site. 

Among them, we have selected: “Navigational characteristics of the Web site;” “Structure” and 

“Effectiveness of its content;” “Informativeness” and “Entertainment.” 

Liu and Arnett (2000) conceptualized Web site quality as the accuracy, completeness, relevancy, ease 

of use, speed, search functionality and organization of the site. They explained that e-commerce success is 

related to four major factors: quality of information (relevant and accurate information), system use, 

195

playfulness (enjoy visiting Web sites, motivate customers to experience participation), and system design 

quality (well organized hyperlinks, high speed of accessing the Web). 

Aladwani and Palvia (2002) defined perceived Web site quality, as an user’s evaluation of a 

Web site’s features that responded to user’s needs and which reflected overall excellence of it. 

They proposed three dimensions of Web site quality: technical adequacy, Web content, and Web 

appearance, and found instead four dimensions. Web content is found to be a bidimensional 

variable with specific content (reflecting concerns related to finding specific details about 

products/services, customer support, privacy policies, other important information) and content 

quality (information usefulness, completeness, and accuracy). 

3.1.1 Navigational characteristics of the Web Site 

Rose et al (1999) hightlighted the importance of download speed, Web interface, search 

functionality, measurement of Web success, security, and Internet standards. The characteristics 

of the products and of the Web sites that are encountered early in online browsing can 

significantly influence the level of arousal and pleasure (emotions) that consumers experience, 

and therefore can influence their responses. Menon and Kahn (2002) show that if the starting 

experiences encountered by potential customers in a simulated Internet shopping trip are high in 

pleasure, then there is a positive influence on approach behaviors (attitudes) and shoppers tend to 

engage in more arousing activities such as more exploration, more tendencies to examine new 

products and stores (= exploratory behavior), and higher response to promotional incentives. 

Further, navigational cues are important in creating or not impeding the flow experienced by 

surfers (Hoffman and Novak, 1996; Novak et al., 2000), which in turn influences formation of a 

positive attitude (Baronas and Louis, 1988; Regan and Fazio, 1977; Eagly and Chaiken, 1993; 

Csikszentmihalyi, 1977). Finally, Lynch, Kent and Srinivasan (2001) show that site quality may 

influence surfers’ probability of buying during the visit and returning to visit the Web site. Thus: 

H1: When consumers surf the Web, navigational characteristics of the Web site are positively related to: 

a) attitude toward the Web site, b) involvement, c) exploratory behavior, d) involvement in 

purchase decisions for the product, and e) purchase intentions. 

3.1.2. Structure 

Huizingh (2000) reported that there are four different navigational structure types: a tree, a tree with a 

return-to-home page button, a tree with some horizontal links and an extensive network. Most of Web 

196

sites have quite a simple structure, over 60% of the Web sites he studied have a tree structure or a tree 

structure supplemented with a back to home page button. 

In the ELM, there are two routes to persuasion: the central route which requires “high effort scrutiny 

of attitude-relevant information,” and the peripheral route which requires less effort to evaluate attitude 

objects. Attitude change is often determined by both processes. Peripheral processes include the use of 

simple decision rules, conditioning processes, and others that do not imply scrutiny of the central merits 

of the attitude object (Petty and Cacioppo, 1986). Among the central cues are: structure of a Web site 

(STR), its informativeness (INFO) and the effectiveness of its information content (EFF). 

Attitudes: In the past, the attitude towards the ad (Aad) was measured by means of likeability-items. 

De Pelsmacker et al. (1998) measured the relationship between Aad and emotional content, information 

content and format characteristics of advertising stimuli. Among the three basic dimensions of Aad, 

structure is one of the two cognitive components. This dimension mostly explain the attitude towards the 

brand which, in turn, is positively correlated with purchase intention (De Pelsmacker, Dedock, and 

Geuens, 1998). 

Involvement: When motivation, ability and opportunity to process information are 

high, consumers process via the central route, focusing and elaborating on (central) 

cues relevant to their needs, values and interests (involvement). When they are low or 

lacking, consumers engage in less effortful processing and use peripheral cues. 

Traditional and modified ELM differentiated by the degree of activeness and 

consciousness when shoppers process ad messages. The modified ELM for web ad 

has more active and more conscious cognitive processing than the traditional one as 

the exposure to ad messages is totally voluntary. When people are highly-involved and 

do not have any ability to process information, peripheral aspects of ad messages are 

used. However, when these people have the ability to process, they engage in active 

and conscious processing or message-related cognitive thinking. Two factors are 

determinant: initial attitude and argument quality of ad messages. 

Exploratory behavior: We did not find any studies on the impact of the structure of the Web site on 

visitors’ exploratory behavior. However, it seems logical that for people who like scrolling and browsing 

throughout the Internet and various types of sites, the structure of the Web site visited might be important 

and positively influence their behavior. However, it is only a supposition, to be tested later. 

There has been no research reported about a possible relationship between the structure and the 

visitors’ purchase intentions, so we surmise that more complex structures might lead to higher intentions 

to purchase. Thus: 

197

H2: When consumers surf the Web, the structure of the Web site is positively related to: a) 

attitude toward the Web site, b) involvement, c) exploratory behavior, and d) purchase 

intentions. 

3.1.3 Effectiveness of the information content 

Among the criteria established for evaluating the information content of commercial Web pages, 

Dholakia and Rego (1998) cited price or value, components or contents, performance, and quality. Web 

page designers are confronted with the problem of how to design pages to make them popular. There are 

two basic creative strategies: either design the page to be cool with colored backgrounds, Java applets, 

sound and video files, or focus on functional presentation of content instead of on novelty and visual 

appeal. 

Kelsey and Misic (1999) mentioned that content and site are evaluated in terms of currency and 

presentation. Top sites are chosen in terms of their unique functionality, design or both (Kelsey and 

Misic, 1999; Johnson and Misic, 1999). Misic and Johnson (1999) described Web-related criteria as 

finding contact information and main page, speed, uniqueness of functionality, ease of navigation, 

counter, currency, wording, color and style. 

Bell and Tang (1998) studied six specific industry sectors and their Web sites rated highly on ease of 

access, content and structure and poorly in the number of unique features. 

Attitude toward the site. Earlier research on the ad value tended to focus on information (Cox, 1962; 

Ratchford, 1980; Stigler, 1961), information content (Resnik and Stern, 1977; Stern, Krugman and 

Resnik,1981), and how informative the ad is perceived to be (Aaker and Norris, 1982; King et al, 1987). 

Forrester Research indicates in a recent study that high-quality content, ease of use, speed and 

frequency of updating are the four main factors contributing to repeat visits (Numbers, 1999). Education 

is where these Web site characteristics were first described and studied. Siu and Chau (1998) noted that 

students perceive the information content of educational Web sites as important and useful, but believe 

that the Web site should be technically easy to master. These authors labeled technological quality and 

appropriateness as information content, whereas quality and originality and perceived loading speed can 

be termed as navigability and quality. Mechitov, Moshkovich and Underwood (2001) compared academic 

Web sites and studied the characteristics of a university Web site and the students’ perception of its 

overall effectiveness. A good Web site is found to be attractive, well-designed (high quality images, 

animation or 3D-graphics without low downloading times), informative with well-identified links to other 

pages, the necessary information displayed concisely and logically, with highly informative content and 

easy access to it, and inclusion of a variety of Web pages devoted to entertainment. 

However, another research found results at variance with those of previous studies. The information 

content of Web pages, per se, does not appear to attract visitors to the Web sites (Dholakia and Rego, 

1998). Web page popularity, however, is found to be strongly and positively influenced by the renewal 

198

and the updating of the information brought to the site in the last three-month period (Dholakia and Rego, 

1998). From this study emerged the notion that the Web is a dynamic and interactive medium, and online 

consumers are variety-seeking and interested people in staying “current” with the firm's offerings 

(Dholakia and Rego, 1998). The number of other Web pages to which a home page is linked, is also 

found to be a significant determinant of popularity (Dholakia and Rego, 1998). Therefore, they support 

the notion that online consumers “surfing the Web” with no prior search objectives appear to be a 

significant number of the visitors to a commercial Web site (Dholakia and Rego, 1998). 

Abels, White and Hahn (1997) looked at the features able to influence a relatively knowledgeable and 

experienced group of WWW users in surfing Web pages. There studied six criteria: use, content, linkage, 

structure, special feature, and appearance. Respondents preferred useful sites, with current information 

and the structure both intelligible and apparent. They emphasized ease of use with the ability to navigate 

easily through a verbal database or to see an overview of the structure so they could make judgments 

quickly about both content and navigation tools. They reacted negatively to sites with superficial content, 

advertisements, and poor navigational characteristics. Some factors such as content and organization have 

been identified as important by library and information professionals (Rettig, 1996). Ease of navigation 

and the ability to determine the structure easily and quickly have been identified through the interface 

design approach (Shneiderman, 1996). Surfers strongly emphasized content but not appearance. The 

findings showed that content and ease of use are the most important criteria affecting continued use of a 

Web site. 

Other authors try to identify the key and effective dimensions of business-to-consumer sites as 

perceived by online consumers, as there is little documentation on that subject (Ranganathan and 

Ganapathy, 2002). They operationalize a B2C site as a site on the WWW where a product or service can 

be purchased. Usefulness of a B2C site depends not only on the information content, but also on tools 

furnished for the navigation and the evaluation of the usefulness of the information. 

Korgaonkar and Wolin (1999) studied the motivations and concern of Web users by applying the uses 

and gratifications concept to this new medium (Eighmey and McCord, 1998; Mukherji et al., 1998). 

Application of this concept improved understanding of Web usage. Among the factors designed to 

capture the surfers’ gratifications and concerns, figured social escapism motivation (~ entertainment), 

Information motivation (for self-education and information needs), Economic motivation (information for 

learning, educational purposes, shopping and buying motivations). King (1996) confirmed the findings of 

previous authors concerning the requirements of surfers who favor regularly updated, well organized, and 

easy to read Web sites. 

Involvement. As previously seen in the ELM, the effectiveness of its information content figures 

among the central cues. De Pelsmacker, Dedock, and Geuens (1998) mentioned two other cognitive 

dimensions, informativeness (EFF and INFO) and comprehension (STR). These two dimensions can 

199

significantly explain attitude towards the brand which, in turn, is positively related to purchase intention. 

We will study if there is a direct and positive relationship between EFF and PURI, and also the causal 

impact of EFF on involvement in purchase decision, which seems to be a mediator in the relation between 

EFF and PURI. However, Okasaki and Rivas (2002) examined the information content variable on 

Japanese Multinational Corporations homepages and found that some cues identified by Resnik and Stern 

(1977) concerning information content should be present to help consumers make a clever purchasing 

decision. 

Several researchers pointed out that the Internet contained a higher level of 

information, compared to other media, and it is a highly involved medium with a mix of 

electronic and print media (Yoon, 2000; Novak and Hoffman, 1996). Thus, we consider 

the measure of information level as an indicator of the degree of involvement (Yoon, 

2000; Okasaki and Rivas, 2002). From purely utilitarian motives (functional attributes), 

surfers may be highly involved in an ad because of the information related to a subject 

area (type of disease) or toward information or a product/service (OTC drug) (= EFF) 

(Park and Young, 1983, 1986). 

Exploratory behavior. No research has been published on the impact of 

effectiveness of the information content on surfers’ exploratory behavior. When surfers 

find an interesting topic to investigate through the Internet, the quality of the information 

content induces them to scroll and browse in order to get the most complete and 

appropriate information about it. 

Outcomes. Others studies consider characteristics of a Web site such as content that refers to the 

information, features or services offered, whereas design is the way used by marketers to show the 

content to the surfers. More specifically, contents have an important role in influencing the purchase 

decision process of a consumer (in our case, pre-purchase and purchase intentions). Design of a Web site 

must contain three important elements. First, there are the ease of and time taken for surfing, second, the 

pages downloaded (Kelsey and Misic, 1999), and finally, the improvement of its visual appeal. Thus, 

consistent links to each page of the site, use of effective navigation buttons, and an index to this site are 

important to develop when designing a B2C web site. Thus: 

H3: When consumers surf the Web, the effectiveness of the information content of the Web site is 

positively related to: a) attitude toward the Web site, b) involvement, c) exploratory behavior, d) 

involvement in purchase decisions for the product, and e) purchase intentions. 

200

3.1.4. Entertainment 

According to Ducoffe (1996), Informativeness and Entertainment are perceptual 

antecedents of ad value. Media context has an important influence on the ad value, since 

previous studies demonstrated that consumers think that newspapers carry the most informative, 

reliable and believable ad, whereas TV and radio rate lower on these attributes (Bauer and 

Greyser, 1968, Becker, Martino and Towners, 1976). On the other hand, TV ads are the most 

entertaining (Larkin, 1979). 

According to McQuail (1983), the value of entertainment rests in its capability to fulfill 

audience needs for escapism, diversion, aesthetic enjoyment or emotional release. People 

grading Web ads high in value would also tend to develop favorable general attitudes. 

Entertainment would retain an independent and direct impact on overall ad attitudes since both 

of these constructs have affective dimensions that are not captured by ad value. Content 

(informativeness) and form (entertainment), crucial to its effectiveness, are both important 

predictors of ad value (Aaker, Batra and Myers, 1992). For the Web, one value-enhancing 

benefit of its interactive ability is the access it will afford consumers to an ad that is timely, 

relevant and convenient, all crucial determinants of its informativeness (Ducoffe, 1996). 

As stated by Eighmey and McCord (1995, 1997), a commercial Web site is a complex 

assortment of uses and gratifications, among which entertainment or stimulation would be the 

leading dimension, followed by informativeness, involvement with others, personal involvement 

with information content, and executional dimensions related to information overload. Still, 

Eighmey and McCord (1997) showed a consistency with previous uses and gratifications studies 

and studies of TV commercials and Web sites. In fact, effective Web sites correspond to the 

valuable intersection of informativeness and entertainment, as surfers are attracted to information 

that adds value in both form and substance. 

Entertainment factor is similar but not identical to terms that load highly on an entertaining 

dimension often found in raters evaluations of TV commercials. Stern (1990) mentioned that 

consumers who consider an ad as entertaining are more likely to credit the brand with positive 

attributes and are more likely to intend to purchase the brand. 

As previously seen in the ELM, we can name attractive sources (in our case, similar to 

entertainment), music (Chebat et al., 2001), humor, and visuals as peripheral cues (Cho, 1999; Shavitt et 

al., 1994). This affective variable influences attitude towards the brand which impacts purchase intention 

(De Pelsmacker et al., 1998). 

201

Involvement toward an ad based on value-expressive motives (aesthetics) leads to 

affective involvement because it appeals emotionally or aesthetically to them (= ENT) 

(Park and Young, 1983, 1986). Thus: 

H4: When consumers surf the Web, entertainment provided by the Web site is positively related 

to their a) attitude and b) involvement. 

3.1.5. Informativeness 

The kind of information that is more often available in a Web site is specific product 

information (Huizingh, 2000). The perception of the Web site content can be measured by the 

degree to which it is considered to be informative (Huizingh, 2000). An informative site provides 

detailed and specific information with respect to products, the company or other relevant topics 

(Huizingh, 2000). 

The informativeness scale looks like the informative/relevant scales that have emerged in 

previous studies of print ads and TV commercials. The Web site Informativeness factor focuses 

on the site as an interactive provider. Intelligent, resourceful, knowledgeable adjectives fit the 

Web site very well but seem less appropriate to TV commercials. The relevance of 

informativeness is supported by the Market Facts TeleNation survey which found that the No.1 

online use is gathering news or information (Maddox, 1998). 

Based on social psychology theory, Chen et al. (1999, 2002) assume that attitude toward the 

ad (Aad) is an equally useful indicator of site value, but the unidimensionality of Aad does not 

provide a complete explanation of consumers’ ratings of the ad (Pashupati, 1994). Consequently, 

Chen et al. (1999, 2002) developed three scales that better correlate and explain attitudes toward 

a Web site. 

Entertainment and Informativeness scales relate to the overall Attitude toward the site (Ast) 

and to each other. These factors accounted for significant variance in the overall Attitude scale. 

Correlations among entertainment and Informativeness are lower than the correlations of each 

dimension with Ast scale. All the factors score high for Web sites having high Ast scores. 

Though Informativeness scores are generally higher than 70, entertainment scores vary 

substantially. This variation in entertainment scores indicates that when the primary purpose of a 

Web site is to convey information, users score it more on Informativeness and less on 

Entertainment. With medium Ast scores, Entertainment scores are fairly low (E

high score on every dimension to gain a high Ast score. These scales complement each other and 

offer some clues on how to improve Web site design and presentation. All the results came from 

data collected by undergraduate and graduate students from corporate and institutional Web sites 

(Chen et al., 1999). 

Chen et al. (2002) studied some selected web sites with Web development professionals and 

online Web users in order to evaluate the scales linked to the attitude toward a Web site. Web 

user’s interests were dramatically different from students’ interests used in the previous study: 

This new set of sites produced higher correlations. Entertainment, Informativeness and 

Organization reappeared even when site selection was done in terms of a single good/bad site 

evaluation. However, correlations among the three dimensions and Ast changed meaningfully 

with the type of respondents. For Web developers, informativeness was the best predictor 

followed closely by organization, entertainment being far distant from them. Meanwhile, for the 

university-based Web users, Informativeness is also the best predictor with value close to that of 

professionals, followed by entertainment which is seven times higher than the one of 

professionals, and finally organization which is the half of the value found for professionals. 

Thus, a good site will be informative and well-organized but may or not be entertaining (high 

Ast scores), whereas for low Ast scores, variances within Informativeness and Organization are 

relatively large and Entertainment is low. 

Chen et al. (2001) applied these scales to pharmacy sites. As Informativeness is an essential 

factor in these sites, they did not differ much in informativeness value, causing higher correlation 

and less variation than other types of Web sites. Another dimension, i.e., trust, should be added. 

However, they found that Informativeness remained the key factor and Trust did not add any 

predictive value to the dimensions of Ast. Consequently, a separate trust factor is rejected. Ast 

remains reliable and unidimensional across changes in types of Web sites, respondents, and 

methods of administration. The three factors accounted for most of the variance in Ast and these 

three dimensions correlated with each other and with Ast significantly. 

As previously seen in the ELM, informativeness is a central cue that impacts attitude, which 

in turn influences purchase intention (De Pelsmacker et al., 1998). According to Chen (1999), as 

informativeness is closely related to attitude toward a Web site, we expect to find a significant 

relationship between these factors. 

No research has been published on the influence of informativeness on consumers’ 

exploratory behavior. However, we can surmise that informativeness, characterized by the 

203

usefulness and the resourcefulness of the site, is important for improving the browsing and the 

scrolling of surfers. Thus: 

H5: When consumers surf the Web, informativeness of the Web site is positively related to: a) 

their attitude toward the Web site and b) their exploratory behavior. 

3.2. Processing Variables 

3.2.1. Attitudes toward the Web Site (Approach/Avoidance Behavior) 

Aad is an indicator of ad effectiveness. Researchers created scales for ratings print ads and 

TV commercials (Wells, 1964; Wells, Leavitt and McConville, 1970; Leavitt, 1970; Aaker and 

Bruzzone, 1981; Aaker and Stayman, 1990; Moldovan, 1985; Schlinger, 1979; Pashupati, 1994). 

One approach for measuring Aad has also proved useful in applied and academic settings 

(Shimp, 1981; MacKenzie, Lutz and Belch, 1986, Baker and Lutz, 1988, Brown and Stayman, 

1992). 

The first processing and affective component of our model is attitude. Stevenson, Bruner and 

Kumar (2000) showed that attitude toward the Web site is a useful construct in understanding the 

impact of a Web site. Shimp (1981), Batra and Ray (1986), MacKenzie, Lutz and Belch (1986), 

Brown and Stayman (1992) found that attitude toward the ad influences brand attitudes and 

purchase intentions. Extending this to the Web context, attitude toward the site will be an equally 

useful indicator of site value. 

Following Jee and Lee (2002), we assume that Web sites look like and reflect the 

characteristics of traditional ads, and therefore, attitude toward the Web site should lead to 

consequences identical to those found in attitude research. The Flow construct mediates the 

effect of attitude toward the site on consumers’ intentions to revisit the site and to purchase this 

product, but it is not needed to predict consumer intentions (Luna et al., 2002). According to 

some authors, attitude toward the ad is an affective construct, which mediates the influence on 

brand attitudes and purchase intentions (Lutz, McKenzie and Belch, 1983; McKenzie et al., 

1986; Mitchell and Olson, 1981; Shimp, 1981; Homer, 1990). In the same manner, attitude 

toward the Web site has a positive and strong impact on attitude toward the ad, attention to the 

commercial, brand attitudes and purchase intentions according to the advertising hierarchy of 

effects model of Bruner and Kumar (2000). From Bruner and Kumar (2000) and Stevenson 

(2000), attitude toward the web site plays an important role in the hierarchy-of-effects, a model 

in which attitude formation goes through a cognitive stage (exploratory behavior), an affective 

204

stage of liking (attitudes), and a behavioral stage of purchase intention in responding to 

advertising and other persuasive marketing messages (Dodds, 1991). Thus: 

H6: When consumers surf the Web, attitude toward the Web site is positively related to: a) involvement 

in purchase decisions for the product, and b) purchase intentions. 

3.2.2 Behavioral Search Variables 

3.2.2.1. Exploratory Behavior 

Exploratory behavior is defined as “behavior with the sole function of changing the stimulus field” 

(Berlyne, 1963). Theory and empirical studies suggest that a two-factor conceptualization of exploratory 

consumer behavior are most useful: exploratory acquisition of products and exploratory information 

seeking (Baumgartner and Steenkamp, 1996). 

Browsing, which is one of the components of exploratory behavior on the Internet, is performed when 

the surfers do not have a precise knowledge of the information that might be available and are not sure 

whether their requirements can be met or how these requirements may be reached. Browsing can be either 

general or purposeful. “Purposive” browsing occurs when the surfers have fairly specific requirements, 

whereas general browsing may be used as an opportunity for the surfers to fine-tune the perception of 

their requirements or to simply keep themselves up-to-date on the latest changes in a specific field or a 

product type (Rowley, 2000). 

Shoppers’ exploratory behavior, characterized by information-search or exploration through 

purchasing, positively influences their attitudes toward the Web site. The more they tend to explore the 

various possibilities offered by the Web, the more they will fine-tune their requirements and have a 

positive idea of the site they visit when surfing the Web, triggering approach behavior toward the Web 

site. Thus: 

H7: When consumers surf the Web, exploratory behavior is positively related to: a) attitude, b) 

involvement, c) involvement in purchase decisions for the product, and d) purchase intentions. 

205

3.2.2.2. Involvement 

The final component of the processing part of our model is involvement. Involvement is a very 

important variable in audience processing of both traditional advertising (Petty and Cacioppo, 1981, 

1983a, 1986) and Web advertising (Raman and Leckenby, 1998; Cho, 1999). Day, Stafford and Camacho 

(1995) refer to involvement as a motivational state influenced by a person’s perception of the object’s 

relevance based on inherent needs, values and interests (Zaichkowsky, 1985). Its major antecedents are 

the characteristics of the person, the stimulus/object, and the situation (Bloch and Richins, 1983; 

Zaichkowsky, 1986). 

We suggest that involved surfers are more prone to search for more information when surfing the pharmaceutical Web sites and in doing 

that, explore more new stimuli and situations because of a higher need for environmental stimulation. Further, Balabanis and Reynolds 

(2001) confirm the influence of live brand attitudes on the attitude formation of online shoppers. Harvin (2000) also indicates that consumers 

are more comfortable with companies’ strong off-line brands that they already know and trust. Yoo and Stout (2001) posited that consumers 

with a high level of product involvement have more intentions to interact with a Web site, leading to more extensive search and more 

interactive functions tried. Finally, highly-involved people will search for more information before purchase, process relevant information in 

depth, and use more criteria in their purchase decisions than others (Leong, 1993; Laaksonen, 1994; Maheswaran and Meyers-Levy, 1990). 

Internet-involved customers will more likely purchase online than people with low levels of involvement (Kwak, Fox and Zinkhan, 2002). 

Therefore: 

H8: When consumers surf the Web, involvement is positively related to: a) attitude b) involvement in purchase decisions for the product, and c) 

purchase intentions. 

3.3.3. Outcomes 

3.3.3.1. Involvement in Purchase Decisions (or Pre-purchase Intentions) 

A consumer buying process is a sequence of several stages: 1-2/ information’s search and evaluation are two important preparatory steps. 

B2C sites which offer navigational tools that would ease the search process are likely to be more effective. 3/ evaluation of alternatives 

before making a decision. Decision aids have a favorable effect on the quality of online purchase decisions (Ranganathan and Ganapathy, 

2002). 

In general, customers often engage in ongoing information collection without specific needs or 

purchase decisions (Block et al. 1986). For them, value-added information can be interesting and helpful. 

Whether the URL to visit comes from word-by-mouth (e.g., reviews of books by other customers or stock 

recommendations by analysts), customers can enjoy this wealth of information about products and 

services not always available in the physical world (Koufaris, 2002). Access to such information can 

improve consumer decision and can be an important incentive for people to shop online (Jarvenpaa and 

Todd, 1997). 

In the traditional hierarchy-of-effects models, product purchase is the ultimate stage of the 

communication process and purchase usually takes place long after exposure to ad messages. Ad 

effects occur not in the short-term, but rather in the long-term, with a series of steps starting from 

unawareness of a particular brand to the actual buying of it. However, with Internet, that 

purchase can probably take place at the same time as their exposure to ad messages or within a 

206

short period of time because shoppers can place an order or request additional information 

instantly and directly via Internet rather than having to order through another method (Cho, 

1999). 

As involvement in purchase decision has been conceptualized as pre-purchase intentions, we can propose that in the hierarchy-of-effects 

model, this factor can considered as a step before purchase intention with a conative dimension: 

H9: When consumers surf the Web, involvement in purchase decisions is positively related to purchase intentions. 

The model we propose in this study is shown in figure 2. To facilitate its comprehension, we 

did not drawn either the non-significant paths or those whose p is less than 0.10. 

[insert figure 2 around here] 

4. Data and Methodology 

We test the structure of the proposed model in the context of a pharmaceutical Web site. 

Health care is one sector of the economy where the Internet has become an invaluable tool for 

communication. For wired consumers, healthcare ranks as the fourth most popular topic on the 

Web (Bellman, Lohse and Johnson, 1999), while among women and seniors, health sites are the 

second most popular destination (Cyber Dialogue, 1999). Obviously, the objectives of 

pharmaceutical companies are geared more towards influencing consumer attitude, rather than 

immediate online purchasing. As we mentioned in the introduction, the majority of the research 

to date is on online retailing, but the empirical context is justified by our objective of developing 

a comprehensive model of the effectiveness of the Web site through its content and its 

informational design (Internet atmospherics), which are represented by navigational 

characteristics, entertainment, informativeness, effectiveness and structure of the Web site. 

The data for the study were collected from the homepage of an OTC drug from one of the 

largest, well-known pharmaceutical companies in North America. (A recent study by Ipsos 

PharmTrends reported that this OTC drug was number two with a 23% U.S. market share, 

compared to the leader’s 32% share). 

Here, we propose a Web-based methodology that minimizes biases of existing ad testing 

methods. Our methodology is based on an experimental method that is totally clear and 

unobtrusive to each respondent. We measured their answers which reflect their flow and viewing 

behavior on a given Web site after they did their normal surfing activities on it. Hence, 

measurements are based on actual respondent behavior and do not suffer from any experimental 

effects or any self-report biases. The methodology is designed to test the effectiveness of 

207

environmental cues on purchase behavior on the Web and is implemented on the Web itself. The 

participants are Web surfers engaged in their own Web surfing activities but are part of a random 

sample of individuals who might or might not be involved with the product featured in the site. 

Thus, the method does not suffer from any biases due to lack of external validity as content is 

viewed in its actual form, by actual viewers, and within its actual viewing environment (within 

the Web site environment for which it is designed). The electronic data is collected from 

respondents instantaneously as they answer the questionnaire, in real time. Thus, the 

methodology potentially provides instantaneous measures of site environmental cues 

effectiveness. 

The questionnaire was a structured, non-disguised instrument, with closed-ended questions to indicate the respondents’ degree of 

agreement/disagreement on a five-point Likert scale, except for involvement which uses a five point-semantic differential scale. It contained 

variables measuring Navigational Characteristics, Entertainment, Informativeness, Effectiveness and Structure of the site that could influence 

the cognitive (i.e., exploratory behavior and involvement), affective (i.e., attitudes toward the site) variables, and outcomes for the OTC drug 

(i.e., involvement in purchase decision and purchase intentions). Finally, some demographic variables were measured. 

All these dimensions have been studied previously, providing a large pool of existing valid items to use in our survey. The most appropriate 

measures for each concept were selected from the literature and adapted to meet the needs of our study. The scale for “Navigational 

characteristics of the Web site” (CHPS), developed by Bell and Tang (1998), was operationalized with eleven items and adapted to our study. 

“Entertainment” (ENT) and “Informativeness” (INFO) of the Web site came from Chen and Wells (1999) and contained respectively two, 

four and three items. “Structure” (STR) of the Web site and “Effectiveness of the information content” (EFFI) of the Web site were created 

from the variables analyzed by Bell and Tang (1998). “Attitude toward the Web site” (ATTI) is described on one hand by the Chen and Wells 

(1999) scale and on another hand, by the Eighmey and McCord (1995) scale as this variable is also considered as a measure of 

approach/avoidance behavior. The “Exploratory Behavior” (EXPB) scale, developed by Novak et al (1997, 2000) was operationalized with 

eight items. The “Involvement” (INV) scale, represented by perceived message relevance and developed by Muehling, Stoltman, and 

Grossbart (1990) and Zaichkowsky (1990) was operationalized with eight items. The scale developed by Gore et al. (1994) about 

involvement in purchase decisions for nonprescription drugs was conceptualized as “Pre-purchase Intentions” (PPURI) and was 

operationalized with seven items. Finally, the last variable, named purchase intentions (PURI), consisted of one indicator. 

5. Results 

5.1 Exploratory Factor Analysis 

Prior to testing the full latent model, we begin with an Exploratory Factor Analysis (EFA) in order to determine how the observed variables 

are linked to their underlying factors. Knowing that some constructs could have items loading on more than one factor, we used a data 

reduction technique to identify the minimal number of factors underlying the observed variables (items) and explaining most of the variance 

observed. Maximum Likelihood Extraction removes highly correlated variables from the data and varimax rotation ascertains that constructs 

are distinct. 

Initially, the number of factors for each construct was equal to 1 except for CHPS, EFF, ATTI and EXPB. All Cronbach alpha coefficients 

were acceptable (higher than 0.60). A preliminary analysis of the psychometric properties of the items composing the different scales 

resulted in deleting certain items presenting poor psychometric properties or changes in Cronbach alpha coefficients. After deleting these 

items, each construct proved to be unidimensional and factorially distinct, all items used to operationalize a specific construct loaded on a 

single factor. The percentage of variance varied between 53.9% for PPURI to 86.5% for INFO. The EFA highlighted the existence of 9 

factors with eigenvalues greater than 1.0. The criterion used to identify and interpret factors was that each item should have a factor loading 

208

greater than 0.4. According to Nunnally (1967), acceptable Cronbach’s α coefficients start at 0.60. In our study, all Cronbach’s coefficients 

except one were greater than 0.70, which indicates very high reliability. The exceptions are the EFF scale with a coefficient α of 0.63. 

Table 1 lists the scale items for each factor, factor loadings, Cronbach’s α and percentage of variance for each factor. 

5.2 Confirmatory Factor Analysis 

[Insert table 1 around here] 

Following Byrne (1994), confirmatory factor analysis (CFA) was used to confirm the 

measurement model before conducting the test of the structural model. The Lagrange Multiplier 

test identified a small number of covariances which were taken into account. The nine-factor 

structure obtained in the EFA was confirmed with a first-order CFA. Estimation of the CFA 

model generated a χ 2 based on 631 degrees of freedom, Satorra-Bentler scaled χ 2 , comparative fit 

indices (CFI) and standardized root mean-square error of approximation (RMSEA) values of, 

respectively, 1087.5, (χ 2 /df=1.72), 894.3, (χ 2 /df=1.42), 0.90 and 0.05. According to Hu and 

Bentler’s (1999) cutoff criteria, the model demonstrated a good fit, taking into account the very 

large number of items and factors analyzed (Baumgartner and Homburg, 1996). 

5.3 Full Structural Model 

The results of the full structural model, along with corresponding fit indices and standardized 

parameter estimates are depicted in table 2. Overall, the findings show strong support for the 

model fit with CFI value of 0.92 and RMSEA statistics of 0.048. The average off-diagonal 

values of the standardized residual matrix was 0.084. As per Hu and Bentler’s (1999) cutoff 

criteria, the fit of this full model is judged acceptable. 


Following Byrne (1994), we then test the significance of individual parameters to assess their 

fit. The results of the factor loadings, and the test statistics indicate that all the factor loadings are 

significant. Finally, we analyzed the path coefficients representing the hypothesized 

relationships between the various constructs. Table 3 and figure 2 provide the regression 

coefficient estimates and their statistical significance. It also shows the standardized values of 

the regression coefficients and relates the paths to our original hypotheses. Twelve subsections 

of hypotheses have non-significant paths. All the others are significant with t tests varying from 

1.80 to 6.11. Concerning the standardized estimates, low values mean that the correlation 

coefficient between two theoretical constructs is not significantly different from zero, making 

some paths non-significant. 

209

6. Test of Hypotheses 


We now briefly discuss the support for individual hypotheses. The results show a strong support for 

the set of hypotheses 1, which are related to the navigational characteristics of the Web site. Specifically, 

there is a positive and very significant relationship of navigational characteristics with surfers’ 

involvement, exploratory behavior, involvement in purchase decisions, whereas it has a weakly positive 

link with attitude toward the Web site. However, hypothesis 1e, related to surfers’ purchase intentions, is 

not supported. 

Similarly we find a partial and weak support for the set of hypotheses 2. Hypotheses 2a and 2c do not 

show any link between structure of the Web site and surfers’ attitude and exploratory behavior towards it. 

However, the other hypotheses (H2b and 2d) show positive and very significant paths between structure 

of the Web site and shoppers’ involvement and pre-purchase intentions. 

The set of hypotheses 3 on the impact of atmospherics clues (information content effectiveness) on 

behavioral variables is mostly supported. H3a representing the effect of EFF on attitude is not supported. 

However, hypotheses 3b, 3c and 3e showed a significant effect of the EFF on INV, EXPB and PURI, 

while EFF (H3d) is significantly and positively related to surfers’ involvement in purchase decision when 

a one sided-test is used (t = 1.40, p < 0.1). 

We find that hypotheses 4a and 4b are supported with a positive and significant link between 

entertainment with the Web site and surfers’ attitudes (one-sided test: t = 1.89, p < 0.05) and involvement. 

Concerning informativeness of the Web site, hypothesis 5a showed significant but negative link of 

informativeness with shoppers’ attitude, whereas for H5b, informativeness has a positive and significant 

influence on exploratory behavior. 

For H6a, there is a positive and very significant path between surfers’ attitude and involvement in 

purchase decisions. For H6b, there is also a positive and significant path between attitude and purchase 

intentions when a one sided-test is used (t = 1.88, p < 0.05). This finding is consistent with expectations 

from prior research. 

The set of hypotheses 7 is partly supported. On one hand, there is no relationship between surfers’ 

exploratory behavior and attitude, involvement and involvement in purchase decisions (7a, 7b and 7c). On 

the other hand, there is a negative and very significant link between exploratory behavior and purchase 

behavior (7d). This finding is interesting in the context of the hierarchy of the effects model. 

A curious result is found for the set of hypotheses 8 related to consumers’ involvement. Contrary to 

previous research, we found no relationship between involvement and involvement in purchase decisions 

(H8b), and involvement and purchase intentions (H8c), and a barely insignificant relationship between 

involvement and attitude (H8a) (the standardized estimate is high, but so is the standard error). 

210

Finally, for the last hypothesis, H9, we found that surfers’ involvement in purchase decisions is 

strongly and positively related to their purchase decisions. 

Overall, the results are very encouraging, with full or partial support for all except one set of 

hypotheses. This study was exploratory as no specific information was published before on most 

of the paths studied in this paper. However, close to 65% of the individual relationships 

proposed in our model were supported, and the overall model itself was found to fit the observed 

data well. 

7. Interpretation and Discussion of Results 

8. 

We will now discuss some of the major findings presented in the previous section, and follow it with 

the theoretical and managerial implications of the study. 

One of the more interesting and useful findings relates to the support for the impact of 

navigational characteristics effectiveness. These cues had impact on attitude toward the Web 

site. We find that navigational characteristics are positively related to surfers’ involvement in 

purchase decision. We could perhaps infer that when navigational characteristics are positive, 

surfers can develop some arousal early in their online browsing (arousal was not measured in the 

study), which can make shoppers more involved in the Web site and more keen to search for 

information (exploratory behavior), affecting positively their involvement in purchase decisions, 

without implicating a concrete purchase of the OTC drug, whatever the channels used 

(significant hypothesis with a one-sided test: t = 1.61; p < 0.1). 

Even though the literature seems to link STR (H2a) and EFF (H3a) to attitude, we did not find such 

results. First, as attitude correlates with involvement pretty well, and as the same inducers of the 

relationship with involvement were statistically significant, we can infer that the effects of involvement 

are already included in the attitude variable. Informativeness showed a relatively significant but negative 

relationship with attitude. In fact, even though it seems there is updated and not repetitive information 

(EFF is positively linked attitude), surfers did not consider the information given by the Web site as 

useful and/or resourceful, because this site seems to them ugly, inducing the development of negative 

affect (avoidance attitude) as they did not like it. 

Central cues such as structure of the Web site (H2b) and effectiveness of its information content 

(H3b) are significantly related to involvement. In fact, the more the STR and EFF are adequate and 

pleasing to the visitors, the more the surfers are engaged in the process of in depth information search. 

Surfers exerted cognitive efforts and used the central route, according to the ELM. Factors such as the 

expertise, credibility, likeability or attractiveness of the source are considered as peripheral cues (Petty, 

211

Cacioppo, Schumann, 1983; Petty and Cacioppo, 1986). Consequently, we can infer that entertainment 

(H4b) is also linked to involvement. In this case, consumers who did not want to give too much attention 

to the content itself, are attracted by visuals and/or attractive sources, giving them the opportunity to 

pursue the visit of the Web site because of the entertainment the site has developed. Visitors are affected 

by the visuals and/or attractive sources, elaborating affective reaction and following the peripheral route. 

Our results support prior research that attitudes toward the Web site have a very strong and positive 

effect on the surfers’ pre-purchase intentions, even though the items concern more specifically the drug 

(or brand of that company).This result is confirmed by previous studies (Shim et al., 2001). However, we 

did not find a very strong relationship between attitude toward the site and purchase intentions, whereas 

research on that subject found the opposite (Stevenson et al, 2000; Notani, 1997). 

Without any other study to support or not our findings, we did not find any significant links between 

consumers’ exploratory behavior and attitude, involvement, and involvement in purchase decisions. 

However, we found a significant but negative path between exploratory behavior and purchase decision. 

The more surfers scroll and browse, the more they are fond of information about the topic they are 

interested in, producing a delay in their purchase intentions. 

One of the interesting results relates to the support for the impact of involvement. We found, perhaps 

surprisingly that involvement has no impact on pre-purchase intentions and purchase behavior. But of 

more interest is the finding that involvement is weakly related to approach attitudes. Normally, highly- 

involved consumers are more attracted by Web site aspects related to the product (information content 

(EFF)), whereas low-involved ones focus more on the peripheral stimuli of the site (ENT) or the site’s 

design characteristics (STR). We can thus infer that highly-involved surfers can develop positive attitudes 

toward the site leading to behaviors such as repeat visits to collect up-to-date information. Obviously, this 

finding carries significant theoretical and managerial implications. Shopper’s involvement has normally a 

positive impact on involvement in purchase decisions. The more involved people are, the more they will 

search for information before purchase, process relevant information in depth and use more criteria in 

their purchase decisions than others. Moreover, Internet-involved consumers will more likely purchase 

online than those with low-levels of involvement. According to our results, this relationship is not 

significant. It is probably due to the fact that involvement and attitude are highly correlated, with the 

effects of involvement being already included in attitude. 

Central cues such as EFF (H3c) and INFO (H5b) impact consumers’ exploratory behavior, whereas 

STR (H2c) does not. EFF and INFO are factors related to the information of the Web site. When the 

manner the information is presented suited visitors and when the site is useful or resourceful to them 

(meaning that surfers find what they want), they live the flow and scroll and browse in order to gather the 

most information possible. According to ELM, these two variables make surfers follow the central route, 

exerting some cognitive efforts. Even though STR is also a central cue, the simple or complex level of 

212

structure does not influence the level of browsing or scrolling of consumers, contrary to our hypothesis. 

They do not seem to exert additional cognitive effort to collect information because the structure is more 

complex. 

Shoppers’ pre-purchase intention is highly and significantly linked to their purchase intentions (H9). 

In fact, it is reasonable for a consumer to be involved in the search for information, before buying a 

product such as an OTC drug. Moreover, according to the hierarchy-of-effects, PPURI can be considered 

as an antecedent of PURI. STR (H2d) and EFF (H3e), two central cues, has a direct impact on PURI. The 

more the structure and information content are complex, the more consumers are engaged in seeking 

information, the more cognitive efforts they develop. EFF also has an indirect path via PPURI toward 

PURI, which is consistent with the hierarchy-of-effects model. ENT, the peripheral cue of the model, has 

an indirect impact on PURI through attitude (affective factor) and PPURI, and through INV (cognitive 

variable) and PPURI. However, INFO impacts directly PURI through either attitude and EXPB. Finally, 

contrary to our hypothesis, EFF did not influence PPURI (H3d). This means that when the information 

content is effective, surfers do not have to engage in further information search before developing or not a 

purchase intent. 

8. Theoretical and Managerial Implications 

The main goal of this study was to examine the impact of atmospherics cues concerning 

Internet advertising on the behavior of consumers and its impact on purchase intentions. We 

defined under the concept of general behavior toward the Web site, several different dimensions. 

First, there is an affective variable represented by attitude toward the site. Second, there are 

behavioral/cognitive variables (because they are related to information search and acquisition) 

such as exploratory behavior and involvement. Finally, there are the outcomes variables, pre- 

purchase intentions and purchase intentions. This research also helps provide a better 

understanding of Internet consumers’ behavior and findings from the study of this model will be 

of interest to behavioral scientists both in retailing and psychology. 

The theoretical implications of this research take several forms. First, our study empirically confirmed 

several relationships discussed in the prior literature. We found that efficient navigational characteristics 

can help develop not only consumers’ involvement and their exploratory behavior (Menon and Kahn, 

2002), but also their attitude toward the site (Baronas and Louis, 1988; Regan and Fazio, 1977; Eagly and 

Chaiken, 1993; Csikszentmihalyi, 1977). The impact of EFF on INV was also found by Yoon (2000). 

According to Okasaki and Rivas 2002, EFF might influence purchase decision (here, purchase intention). 

Consistent with Chen et al. (1999, 2002), ENT and INFO are two other but independent dimensions of 

ATTI with which they correlate well. In our case, ENT and INFO are significantly related to ATTI (De 

213

Pelsmacker et al, 1998). ATTI is linked to PURI as Brown and Stayman (1992) and De Pelsmacker, et al, 

(1998) found. We confirmed the positive relationship between ENT and INV found by Park and Young 

(1983, 1986). 

More importantly, our model tested other relationships that had not been studied previously in the 

literature: the impact of CHPS and EFF, two central cues, on PPURI; STR, EFF and INFO on EXPB, 

STR and PPURI on PURI; EXPB on INV and on PPURI, EXPB on INV, and finally, ATTI on PPURI. In 

summary, ENT--> INV and ATTI; STR--> INV; INFO--> ATTI and EXPB; EFF--> INV and EXPB; 

CHPS--> INV, ATTI and EXPB. 

We adapted the ELM to our study. In that case, as a consumer’s motivation (represented by EXPB) 

using central route to process a site increases, the impact of central processing on attitudes toward the site 

should increase, the impact of peripheral processing should decrease, and the impact of attitudes toward 

the site on purchase intention should increase (Petty and Cacioppo, 1986). Thus, EXPB has its 

moderating effects by changing the central route (increase) that takes place and by changing peripheral 

perceptions (decrease). Consequently, STR and EFF follow the central route via INV (cognitive aspect); 

ENT and CHPS impacts on either INV (cognitive, central route) or ATTI (affective, peripheral route) and 

INFO impacts totally on ATTI (affective variable). Here, EXPB should moderate the relationship 

between EFF and INV, but not between INFO and ATTI. CHPS is linked with both peripheral (ATTI) 

and central (INV) cues, in the presence of EXPB. Here, we believe that the central route (using INV) 

might be preferentially used, compared to the peripheral route (using ATTI). Thus, we can mention that 

INFO, EFF, STR and CHPS (central cues) will take the central route, whereas ENT will follow both 

routes with a predilection for the affective route. 

Further, our study fails to find support for some relationships examined in previous studies, 

emphasizing the need for further research in this area. For instance, while CHPS was found to positively 

influences PURI (Lynch, Kent and Srinivasan, 2001), our study showed no significant direct path 

between both variables. Before developing a purchase intention, surfers have to navigate through the site 

or find more information via other channels. Whereas STR and ATTI and EFF and ATTI are not 

significant linked in our research, De Pelsmacker et al (1998) reported significant links. Our problem 

probably resides in the high correlation between ATTI and INV. Contrary to Kwak (2002), there is no 

significant direct link between INV and PURI. In the same way, INV and PPURI (Leong, 1993) and INV 

and ATTI (Yoo and Stout, 2001) do not display any significant relationship in our research because of the 

correlation of INV and ATTI. 

Finally, we note that some of our results contradicted previous findings, highlighting the need for 

more empirical research to increase our knowledge in this area. For instance, EXPB did not positively 

influence PURI, but instead negatively, and the positive relationship between STR and INV is reversed. 

With high involvement, but no ability to process information, peripheral cues of ad messages are used in 

214

order to develop involvement toward the brand or product. However, with high involvement and ability to 

process information, central cues are used. 

Finally, Bruner and Kumar (2000) found that attitudes toward the Web site impact purchase 

intentions. On the other hand, Kwak et al. (2002) found that attitudes toward online advertisements 

generally did not impact the overall Internet purchase process. We found the same results as Kwak et al. 

(2002), but with the effect of attitudes toward the site on pre-purchase intentions and purchase intentions 

for the product. 

This study provides marketing practitioners with insights into some of the individual and 

behavioral variables that influence consumers’ purchase intentions for an OTC drug when they 

use the Internet channel to seek and collect information about it. 

The findings generally indicate what types of navigational cues (CHPS), which kinds of design and 

effective information content (INFO, EFF and STR) are more likely to involve surfers in seeking product- 

related information from the Internet. 

Given the early positioning of information search in the buying decision process, if marketers can 

identify which consumer segments in their market niche rely more heavily on navigational characteristics, 

information content and structure of the Web site (central cues) or on entertainment (peripheral cue), how 

to decrease the difficulty of navigating on their site, and how to build a site with a structure and 

information convenient and in the same time appealing to their needs, they can tailor their communication 

strategies to better suit these segments. 

The study is not without limitations. Empirical surveys gathered on the Internet may have questions 

about external validity, more so since our sample of respondents was gathered from one single Web site 

and was composed of students with computer and Internet experience. The addition of emotions (pleasure 

and arousal) might have explained some paths considered in our study as non significant or through 

indirect paths. 

Some other areas for future research can be suggested here. It will have been interesting to also 

measure and add to our model attitude toward the brand in order to have a direct link with PPURI and 

PURI related to the brand. Knowing how consumers consider an OTC drug (search or experience 

product) will help us in testing the ELM in a web context. Other additional variables might be trust and 

experience as they influence consumers’ attitude and purchase intention at a specific Web site. Metrics for 

evaluating our environmental cues such as stickiness (average time per visit, frequency and recency) 

(Bhat et al., 2002) might also be studied in order to confirm our model. 

Researchers could use cross-cultural research to find out if people from different populations 

and cultures differ in their behavior vis-à-vis our model. Longitudinal studies could be done to 

trace the evolution and adaptation of consumer behavior when technological developments and 

215

improvements are brought into navigational characteristics, adding visual and audio capabilities 

such as voice intensity, intonation and speech rate (Gélinas-Chebat et Chebat, 1992; Gélinas- 

Chebat, Chebat and Boivin, 1996) and improving the quality of the information found on the 

Web. 

References 

Aaker, DA and Bruzzone, DE. Viewer Perceptions of Prime-Time Television Advertising. Journal of Advertising 

Research 21(5) (1985): 15-23. 

Aaker, D.S. and Norris, D.: Characteristics of TV Commercials Perceived as Informative. Journal of Advertising 

Research 22 (2) (1982): 61-70. 

Aaker, D.A. and Stayman, DM. Measuring Audience Perceptions of Commercials and Relating them to Ad Impact. 

Journal of Advertising Research 30(4) (1990): 7-17. 

Abels, E.G., White, M.D., and Hahn, K. Identifying User-Based Criteria for Web Pages. Internet Research 7 (4) 

(1997): 252-62. 

Aladwani, A.M. and Palvia, P.C. Developing and Validating an Instrument for Measuring User-Perceived Web 

Quality. Information & Management 39 (2002): 467-76. 

Anderson, PM. Personality, Perception and Emotional-state Factors in Approach-Avoidance Behaviors in the Store 

Environment. AMA Educators’ Proceedings, 35-39, 1986. 

Baker, J. The Role of Environment in Marketing Services: The Consumer Perspective, in The Services marketing 

Challenges Integrated for Competitive Advantage, J.A. Czpeil and C. Congram and J. Shanaham (Eds.). Chicago, 

IL: AMA, 1986. 

Baker, J., D. Grewal, and Parasuraman, S. The Influence of Store Environment on Quality Inferences and Store 

Image. Journal of Academy Marketing Science, 22 (Autumn) (1994): 328-39. 

Baker, WE. and Lutz, RJ. The Relevance-Accessibility Model of Advertising Effectiveness. In: S. Hecker and DW 

Stewart, eds. Nonverbal Communication in Advertising. Lexington, MA : Lexington Books, 1988. 

Balabanis, G. and Reynolds, N.L.: Consumer Attitudes towards Multi-Channels Retailers' Web Sites: The Role of 

Involvement, Brand Attitude, Internet Knowledge and Visit Duration. Journal of Business Strategies 18(2) (2001): 

105-132. 

Baronas, A.K. and Louis, M.R.: Restoring a Sense of Control during Implementation. MIS Quarterly 12 (1988): 

111-123. 

Batra, R. and Ray, M.L.: Affective Responses Mediating Acceptance of Advertising. Journal of Consumer 

Research 13(2) (1986): 234-249. 

Bauer, R.A. and Greyser, S.A. Advertising in America Consumer View. Boston, MA: Harvard University, 1968. 

Baumgartner, H. and Homburg, C.: Applications of Structural Equation Modeling in Marketing and Consumer 

Research: A Review. International Journal of Research in Marketing 13 (1996): 139-161. 

Baumgartner, H. and Steenkamp, J.B.E.M.: Exploratory Consumer Behavior: Conceptualization and Measurement. 

International Journal of Research in Marketing 13(2) (1996): 121-137. 

216

Becker, L.B., Martino, R.A., and Towners, W.M.: Media Advertising credibility. Journalism Quarterly, 53 (1976) : 

216-22. 

Bell, H. and Tang, N.K.H.: The Effectiveness of Commercial Internet Web Sites: A User’s Perspective. Internet 

Research 8(3) (1998): 219-228. 

Bellman, S., Lohse, G.L., and Johnson, E.J.: Predictors of Online Buying Behavior. Communications of the ACM 

42(12) (1999): 32-38. 

Berlyne, D.E.: Motivational Problems Raised by Exploratory and Epistemic Behavior, in Psychology: A Study of 

Science 5, S. Koch (ed.), New York: McGraw-Hill, 284-364, 1963. 

Bhat, S, Bevans, M., and Sengupta, S. Measuring Users' Web Activity to Evaluate and Enhance Advertising 

Effectiveness. Journal of Advertising, 31 (3) (2002): 97-106. 

Bloch, P.H. and Richins, M.L.: A Theoretical Model for the Study of Product Importance Perceptions. Journal of 

Marketing 47 (Summer) (1983): 69-81. 

Brännback, M. Is the Internet Changing the Dominant Logic of Marketing? European Management Journal 15 (6) 

(1997): 698-707. 

Brown, S.P. and Stayman, D.M.: Antecedents and Consequences of Attitudes toward the Ad: A Meta Analysis. 

Journal of Consumer Research, 19 (June) (1992): 34-51. 

Bruner II, G.C. and Kumar, A.: Web Commercials and Advertising Hierarchy-of-Effects, Journal of Advertising 

Research 40(1/2) (2000): 35-42. 

Byrne, B.M.: Structural Equation Modeling with EQS and EQS/Windows: Basic Concepts, Applications, and 

Programming, eds. Sage Publications, 1994. 

Cacioppo, J.T. and Petty, R.E.: The Need for Cognition. Journal of Personality and Social Psychology 42 (1982): 

116-131. 

Cacioppo, J.T. and Petty, R.E. Effects of Message Repetition on Argument Processing, Recall, and Persuasion. 

Basic and Applied Social Psychology 10 (1989): 3-12. 

Cacioppo, J.T., Petty, R.E., and Kao, C.F. Central and Peripheral Routes to Persuasion: An Individual Difference 

Perspective. Journal of Personality and Social Psychology 51 (1986): 1032-1043. 

Chebat, J.C, Gelinas-Chebat, C., and Vaillant, D.: Environmental Background Music and in-Store Selling, Journal 

of Business Research 54(2) (2001): 115-123. 

Chen, Q., Clifford, S.J. and Wells, W.D.: Attitude toward the Site II: New Information, Journal of Advertising 

Research 42(2) (2002): 33-45. 

Chen, Q. and Wells, W.D.: Attitude toward the Site. Journal of Advertising Research 39(5) (1999): 27-37. 

217

Cho, C.H.: How Advertising Works on the World Wide Web: Modified Elaboration Likelihood Model. Journal of 

Current Issues and Research in Advertising 21(1) (1999): 33-49. 

Cox, D.F. The Measurement of Information Value, in Emerging Concepts in Marketing. New York: American 

Marketing Association, 1962. 

Csikszentmihalyi, M. Beyond Boredom and Anxiety. M. Csikszentmihalyi, Ed (second ed.). San Francisco, 

Washington, London: Jossey-Bass, 1977. 

Dailey, L. (forthcoming), Navigational Web Atmospherics Explaining the Influence of Restrictive Navigation Cues. 

Journal of Business Research. 

Darden, W.R., Erdem, O., and Darden, D.K.: A Comparison and Test of Three Causal Models of Patronage 

Intentions. In: William, RD, Robert, FL, eds. Patronage Behavior and Retail Management. New York (NY): North 

Holland, 1983, 29-43. 

Day, E., Stafford, M.R. and Camacho, A.: Research note: Opportunities for involvement research--A scaledevelopment 

approach. Journal of Advertising 24(3) (1995): 69-75. 

De Pelsmacker, P, B Dedock, and M Geuens. A Study of 100 Likeable TV Commercials: Advertising 

Characteristics and the Attitude towards the Ad. Marketing and Research Today 27 (4) (1998): 166-179. 

Dholakia, U.M. and Rego, L.L. What Makes Commercial Web Pages Popular ? European Journal of Marketing 32 

(7/8) (1998): 724-36. 

Dodds, WB. In Search of Value: How Price and Store Name Information Influence Buyers’ Product Perception. The 

Journal of Services Marketing. 5 (3) (1991): 27-36. 

Donovan, R.J. and Rossiter, J.R. Store Atmosphere: An Environmental Psychology Approach. Journal of Retailing 

58(1) (1982): 34-57. 

Donovan R.J., Rossiter, J.R., Marcoolyn, G. and Nesdale, A. Store Atmosphere and Purchase Behavior. Journal of 

Retailing 70 (1994): 283-294. 

Dowling, G. Perceived Risk: The Concept and its Measurement. Psychology & Marketing 3 (1986): 193-210. 

Ducoffe, R.H.: Advertising Value and Advertising on the Web. Journal of Advertising Research 36(5) (1996): 

21-35. 

Eagly, A.H. and Chaiken, S.: The Psychology of Attitudes. Fort Worth (TX): Harcourt Brace Jovanovich, 1993. 

Eighmey, J.: Profiling User Responses to Commercial Web Sites. Journal of Advertising Research 37(3) (1997): 

59-66. 

Eighmey, J. and McCord, L.: Adding Value in the Information Age: Used and Gratifications of the World Wide 

Web, in Proceedings of the Conference on Telecommunications and Information Markets, R.R. Dholakia and D.R. 

Fortin, eds., Newport: University of Rhode Island, 1995. 

Eroglu, S.A., Machleit, K.A. and Davis, L.M.: Atmospherics Qualities of Online Retailing: A Conceptual Model 

and Implications. Journal of Business Research 50 (2001): 177-184. 

Gélinas-Chebat, C., Chebat, JC, and Boivin, R. Impact of Male and Female Voice Cues on Consumers' Attitudes in 

Telemarketing. 

Gelinas-Chebat, C. and Chebat, J.C. Effects of two Voice Characteristics on the Attitudes toward Advertising 

Messages. The Journal of Social Psychology 132(4) (1992): 447-459. 

218

Gore, P., Madhavan, S., McClung, G., and Riley, D.: Consumer Involvement in Non Prescription Medicine 

Purchase Decisions. Journal of Health Care Marketing 14(2) (1994): 16-20. 

Grossbart, S.L., Mittelstaedt, R.A., Curtis, W.N. and Rogers, R.D.: Environmental Sensitivity and Shopping 

Behavior. Journal of Business Research 3(October) (1975): 281-294. 

Harvin, R.: In Internet Branding, The Off-Lines Have It. Brandweek, 41(4) (Jan. 24, 2000): 30-31. 

Hoffman, D.L. and Novak, T.P.: Marketing in Hypermedia Computer-Mediated Environments: Conceptual 

Foundations. Journal of Marketing 60 (3) (1996): 50-68. 

Homer, P.M.: The Mediating Role of Attitude Toward the Ad: Some Additional Evidence, Journal of Marketing 

Research 27 (February 1990): 78-86. 

Hu, L. and Bentler, P.M.: Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria 

Versus New Alternatives. Structural Equation Modeling 6(1)(1999): 1-55. 

Huizingh, E.K. The Content and Design of Web Sites: An Empirical Study. Information & Management 37 (3) 

(2000): 123-34. 

Jarvenpaa, S. and Todd, P. Consumer Reactions to Electronic Shopping on the World Wide Web). International 

Journal of Electronic Commerce 1(2) (1997): 59-88. 

Jee, J. and Lee, W.N.: Antecedents and Consequences of Perceived Interactivity: An Exploratory Study. Journal of 

Interactive Advertising 3(1): 1-16 (2002). (jiad.org/vol3/no1/jee). 

Johnson, K.L. and Misic, M.M.: Benchmarking: A Tool for Web site Evaluation and Improvement. Internet 

Research 9(5) (1999): 383-392. 

King, A.B. What Makes a Great Web Site ? (1996). 

Webreference.com, Webreference.com/greatsite.html. 

King, W.R., Whitehall, K., Reid, L.N., Tinkham, S.F. and Pokrywczynski, J. The Perceived Informativeness of 

National and Retail Advertising. Current Issues and Research in Advertising 10 (1) (1987): 173-198. 

Korgaonkar, P.K. and Wolin, L.D.: A Multivariate Analysis of Web Usage. Journal of Advertising Research 

(1999): 53-68. 

Kotler, P.: Atmosphere as a Marketing Tool. Journal of Retailing 49(winter) (1973): 48-64. 

Koufaris, M. Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior. 

Information Systems Research 13(2) (2002): 205-224. 

Kwak, H., Fox, R.J. and Zinkhan, G.M.: What Products Can Be Successfully Promoted and Sold via the Internet? 

Journal of Advertising Research 42(1) (2002): 23-38. 

Laaksonen, P.: Consumer Involvement: Concepts and Research. London: Routledge, 1984. 

Larkin, E.F. Consumer Perceptions of the Media and their Advertising Content. Journal of Advertising 8 (2) (1979): 

5-7. 

Leavitt, C. A Multidimensional Set of Rating Scales for Television Commercials. Journal of Applied Psychology 

54(5) (1970): 427-429. 

Leong, S.M.: Consumer Decision Making for Common, Repeat-Purchase Products: A Dual Replication. Journal of 

Consumer Psychology 2(2) (1993): 193-208. 

219

Liu, C. and Arnett, K.P. Exploring the Factors Associated with Web Site Success in the Context of Electronic 

Commerce. Information & Management 38 (1) (2000): 23-33. 

Luna, D., Peracchio, L.A. and de Juan, M.: Cross-Cultural and Cognitive Aspects of Web Site Navigation, Journal 

of the Academy of Marketing Science 30(4) 2002: 397-410. 

Lutz, R.J., MacKenzie, S.B. and Belch, G.E.: Attitude toward the Ad as a Mediator of Advertising Effectiveness: 

Determinants and Consequences, in Advances in Consumer Research 10, Bagozzi R.P. and Tybout A.M. eds., Ann 

Arbor, MI: Association for Consumer Research, 1983: 532-539. 

Lynch, P.D., Kent, R.J. and Srinivasan, S.S.: The Global Internet Shopper: Evidence from Shopping Tasks in 

Twelve Countries. Journal of Advertising Research 41(3) (2001): 15-23. 

MacInnis, D.J. and Jaworski, B.J.: Information Processing from Advertisements: Toward an Integrative Framework. 

Journal of Marketing, 53 (October 1989): 1-23. 

MacKenzie, S.B. Lutz, R.J. and Belch, G.F.: The Role of Attitude toward the Ad as a Mediator of Advertising 

Effectiveness: A Test of Competing Explanations, Journal of Marketing Research 23(2) (1986): 130-143. 

Maheswaran, D. and Meyers-Levy, J.: The Influence of Message Framing and Involvement. Journal of Marketing 

Research 27 (August 1990): 361-367. 

McGaughey, R.E. and Mason, K.H.: The Internet as a Marketing Tool. Journal of Marketing Theory and Practice 

6(3) (1998): 1-11. 

McQuail, D. Mass Communication Theory: An Introduction. London: Sage, 1983. 

Mechitov, A.I., Moshkovich, H.M., Underwood, S.H. and Taylor, R.G. Comparative Analysis of Academic Web 

Sites. Education 121 (4) (2001): 652-62. 

Mehrabian, A. and Russell, J.A.: The Basic Emotional Impact of Environments. Perceptual and Motor Skills 38 

(1974): 283-301. 

Menon, S. and Kahn; B.: Cross-Category Effects of Induced Arousal and Pleasure on the Internet Shopping 

Experience. Journal of Retailing 78(1) (2002): 31-40. 

Miller, T.E. and Reents, S. The Health-Care Industry in Transition: The Online Mandate to Change, in American 

Internet User Survey: Cyber Dialog. 

Milliman, R.E. and Fugate, D.L. Atmospherics as an Emerging Influence in the Design of Exchange Environments. 

Journal of Marketing Management 3 (Spring/Summer) (1993): 66-74. 

220

Misic, M.M. and Johnson, K. Benchmarking: A Tool for Web Site Evaluation and Improvement. Internet Research 

9 (5) (1999): 383-92. 

Mitchell, A.A. and Olson, J.C.: Are Product Attribute Beliefs the Only Mediator of Advertising Effects on Brand 

Attitude? Journal of Marketing Research 18 (August 1981): 318-332. 

Moldovan, SE. Copy Factors Related to Persuasion Scores. Journal of Advertising Research 24(6) (1984): 16-22. 

Muehling, D.D., Stoltman, J.J. and Grossbart, S.L.: The Impact of Comparative Advertising on Levels of Message 

Involvement. Journal of Advertising 19(4) (1990): 41-50. 

Mukherji, J., Mukherji, A. and Nicovich, S. Understanding Dependency and Use of the Internet: A Uses and 

Gratifications Perspective. Boston, MA, 1998. 

Notani, AS. Perceptions of Affordability: Their Role in Predicting Purchase Intent and Purchase. Journal of 

Economic Psychology 18(5) (1997): 525-546. 

Novak, T.P., Hoffman, D.L. and Yung, Y.F. Modeling the Flow Construct in Online Environments: A Structural 

Modeling Approach. Marketing Science 19 (1) (2000): 22-42. 

Numbers. Business 2.0 (April 1999), 108. 

Nunnally, J.C.: Psychometric Theory. New York: McGraw-Hill, 1978. 

Obermiller, C. and Bitner, M.J.: Store Atmosphere: A Peripheral Cue for Product Evaluation, in American 

Psychological Association Annual Conference Proceedings, D.C. Stewart (Ed.) Vol. 52-53: American 

Psychological Association, 1984. 

Okasaki, S. and Rivas, J.A. A Content Analysis of Multinational's Web Communication Strategies: Cross-Cultural 

Research Framework and Pre-testing. Internet Research 12 (5) (2002): 380-390. 

Park, CW and Young, SM. Consumer Response to Television Commercials: The Impact of Involvement and 

Background Music on Brand Attitude Formation. Journal of Marketing Research 23 (February) (1986): 11-24. 

Petty, R.E., Cacioppo, J.T.: Attitudes and Persuasion: Classic and Contemporary Approaches. Dubuque, IA: 

William Brown Co, 1983a. 

Petty, R.E. and Cacioppo, J.T.: Communication and Persuasion: Central and Peripheral Routes to Attitudes 

Change. New York: Springer-Verlag, 1986. 

Petty, R.E., Cacioppo, J.T. and Schumann, D.: Central and Peripheral Routes to Advertising Effectiveness: The 

Moderating Role of Involvement. Journal of Consumer Research 10(2) (1983): 135-146. 

Petty, R.E., J.T. Cacioppo, and D. Schumann (1983), “Central and Peripheral Routes to Advertising Effectiveness: 

The Moderating Role of Involvement,” Journal of Consumer Research, 10 (2), 135-46. 

Ratchford, B.T.: The Value of Information for Selected Appliances. Journal of Marketing Research 27 (1) (1980): 

14-25. 

Regan, D.T. and Fazio, R.H.: On the Consistency between Attitudes and Behavior. Journal of Experimental 

Psychology 13 (1977): 28-45. 

Resnik, A. and Stern, B.L. An Analysis of the Information Content of Television Advertising. Journal of Marketing 

41 (1) (1977): 50-53. 

Rettig, J. “Beyond 'Cool'-Analog Models for Reviewing Digital Resources, Online 20 (5) (1996): 52-64. 

221

Rose, G., Khoo, H. and Straub, D.W. Current Technological Impediments to Business-to-Consumer e-Commerce, 

Communications of the Association for Information Systems, 1 (16) (1999). 

Rosen, DE and Purinton, E. Website Design Viewing the Web as a Cognitive Landscape. Journal of Business 

Research (forthcoming). 

Rowley, J.: Product Search in e-Shopping: A Review and Research Propositions. Journal of Consumer Marketing 

17(1) (2000): 20-35. 

Schlinger, MJ. A Profile of Responses to Commercials. Journal of Advertising Research 19(2) (1979): 37-46. 

Shneiderman, B. Designing Information-Abundant Websites. http://www.cs.umd.edu/projects/hcil//pubs/techreports.shtml, 

1996. 

Shavitt, S, Swan, S., Lowrey, TM, and Wanke, M. The Interaction of Endorser Attractiveness and Involvement in 

Persuasion Depends on the Goal that Guides Message Processing. Journal of Consumer Psychology 3 (1994): 137- 

162. 

Shim, S., Eastlick, M.A., Lotz, S.L. and Warrington, P. An Online Prepurchase Intentions Model: The Role of 

Intention to Search. Journal of Retailing 77 (3) (2001): 397-416. 

Shimp, T.A.: Attitude toward the Ad as a Mediator of Consumer Brand Choice. Journal of Advertising 10(2) 

(1981): 9-15. 

Siu, W.-S. and Chau, L.L.-F. Teaching Marketing Research with the Internet. Journal of Education for Business 

(1998) : 44-49. 

Stern, B.L., Krugman, D.M., and Resnik, A. Magazine Advertising: An Analysis of its Information Content. 

Journal of Advertising Research 21 (2) (1981): 39-44. 

Stevenson, J.S., Bruner II, G.C., and Kumar, A.: Web Page Background and Viewer Attitudes. Journal of 

Advertising Research 40(1/2) (2000): 29-34. 

Stigler, G. The Economics of Information. Journal of Political Economy 69 (1961): 213-25. 

Turley, L.W. and Milliman, R.E. Atmospherics Effects on Shopping Behavior: A Review of the Experimental 

Evidence. Journal of Business Research 49 (2000): 193-211. 

van der Heijden, H., Verhagen, T, and Creemers, M. Dieting Online Purchase Behavior: Replications and Tests of 

Competing Models. Proc. 34th Hawaii Internat. Conf. System Sci, Maui, HI. 2001. 

Venkatraman, M.P. and Price, L.L.: Differentiating between Cognitive and Sensory Innovativeness: Concepts, 

Measurement, and Implications. Journal of Business Research 20(4) (1990): 293-315. 

Yoo, C.Y. and Stout, P.A. Factors Affecting Users' Interactivity with the Web Site and the Consequences of Users' 

Interactivity, in Proceedings of the 2001 Conference of the American Academy of Advertising, C.R. Taylor (Ed.). 

Villanova University, Villanova, PA: American Academy of Advertising, 53-61, (2001). 

Yoon, D. Use of Endorsers in Internet Advertising: A Content Analysis of Top 100 American Advertisers Web 

Pages, in Proceedings of the 1999 Conference of the American Academy of Advertising, M.A. Shaver (Ed.). 

Michigan State University: East Lansing, MI, (2000). 

Zaichkowsky, J.L.: Measuring the Involvement Construct, Journal of Consumer Research 12 (December 1985): 

341-352. 

Zaichkowsky, J.L.: Conceptualizing Involvement, Journal of Advertising 15(2) (1986): 4-14. 

222

Zaichkowsky, J.L. The Personal Involvement Inventory: Reduction, Revision, and Application to Advertising. 

Journal of Advertising 23 (4) (1994): 59-70. 

223

Table 1: Results of Confirmatory Factor Analysis (after Deletions) 

Constructs Items % of 

CHPS 

STR 

EFF 

ENT 

INFO 

Navigational problems are limited 

Easy keywords to find information are used 

Product information is immediately accessible 

There is linkage to sites with relevant information 

It allows a great overview of its structure 

The structure is straightforward 

Information is not repetitive 

Information is superficial 

Information is accurate 

Information is up-to-date 

Exciting site 

Imaginative site 

Entertaining site 

Attractive site 

Useful site 

Resourceful site 

ATTI Surfing this Web site is an excellent way for me to spend my time 

I was smiling while I was exploring this Web site 

I was part of a like-minded group of people while using this Web site 

This Web site is an essential tool 

This Web site was a playful experience 

Compared with other Web sites, I would rate this one as the best 

EXPB When I hear about a new Web site, I'm always eager to check it out 

Surfing the Web to see what’s new is a waste of my time 

I like to browse the Web and find out about the latest sites 

224 

variance 

56.9 

Factor 

loadings 

0.752 

0.721 

0.607 

0.597 

78.9 0.886 

59.3 

67.6 

0.783 

0.509 

0.810 

0.913 

0.805 

0.897 

0.818 

0.690 

0.467 

86.5 0.677 

59.1 

0.468 

0.540 

0.792 

0.731 

0.844 

0.414 

0.614 

0.666 

Cronbach 

alpha 

0.78 

0.72 

0.63 

0.84 

0.84 

0.88

INV Important to me (1)... Unimportant to me (5) 

Means nothing to me (1) Means a lot to me (5) 

Worth remembering (1)... Not worth remembering (5) 

Relevant to my needs (1)... Irrelevant to my needs (5) 

Worth paying attention to (1). Not worth paying attention to (5) 

Convincing (1)... Unconvincing (5) 

Boring (1) Interesting (5) 

PPURI It takes a very long time to decide before buying drugs 

I get as much information as possible before purchasing a drug 

I always compare product characteristics among brands of a specific drug 

Before looking at this site, I will be interested in reading about the needed 

drug 

PURI Extremely willing to buy (1) Definitively not willing to buy (5) 

225 

64.6 0.506 

57.6 

53.9 

0.392 

0.668 

0.488 

0.733 

0.667 

0.588 

0.496 

0.454 

0.667 

0.852 

0.823 

0.707 

0.72 

0.87 

0.71

Table 2: Full Structural Model Fit 

FIT CRITERIA FULL MODEL 

Average Off-Diagonal Absolute Standardized 

Residuals 

0.084 

Distribution of Standardized Residuals Very good 

X 2 

924.4 

S-Bx2 767.5 

Degrees of Freedom (df) 578 

S-Bx2/df 1.33 

NFI 0.82 

NNFI 0.91 

CFI 0.92 

Adjusted CFI 0.94 

RMSEA 0.048 

226

Table 3: Statistical Significance of Regression Coefficients’ Estimates 

H Parameter estimate Standard 

error 

Test 

statistic 

Standardized 

estimate 

1CHPS� ATTI H1a *(one-sided test) 0.068 1.608 0.144 

11CHPS� INV H1b *** 0.039 2.951 0.206 

CHPS� EXPB H1c **** 0.060 3.431 0.325 

CHPS� PPURI H1d **** 0.068 4.569 0.378 

CHPS� PURI H1e NS 0.096 0.534 0.041 

STR� ATTI H2a *(one-sided test) 0.187 0.845 0.136 

STR� INV H2b **** 0.090 5.920 0.619 

STR� EXPB H2c NS 0.121 -1.137 -0.143 

11STR� PURI H2d ** 0.235 2.515 0.307 

EFF� ATTI H3a NS 0.167 0.934 0.104 

EFF� INV H3b ** 0.113 2.192 0.223 

11EFF� EXPB H3c *** 0.167 3.034 0.409 

EFF� PPURI H3d *(one-sided test) 0.150 1.402 0.130 

EFF� PURI H3e **(one-sided test) 0.219 1.805 0.159 

ENT� ATTI H4a **(one-sided test) 0.086 1.889 0.235 

11ENT� INV H4b *** 0.026 3.233 0.163 

1INFO� ATTI H5a * 0.166 -2.083 -0.310 

INFO� EXPB H5b ** 0.078 2.433 0.208 

ATTI�PPURI H6a **** 0.092 4.987 0.423 

ATTI� PURI H6b **(one-sided test) 0.116 1.884 0.132 

EXPB� ATTI H7a NS 0.127 -0.043 -0.005 

227

EXPB� INV H7b NS 0.080 0.062 0.006 

EXPB� PPURI H7c NS 0.121 -1.266 -0.117 

EXPB�PURI H7d *** 0.188 -2.605 -0.244 

INV� ATTI H8a NS 0.230 1.251 0.223 

INV� PPURI H8b NS 0.136 0.596 0.056 

INV� PURI H8c NS 0.283 -0.541 -0.069 

PPURI� PURI H9 **** 0.123 4.959 0.398 

**** significant at p < 0.001 

*** significant at p < 0.01 

** significant at p < 0.05 

* significant at p < 0.10 

228

Figure 1: Initial Conceptual Model 

RESPONSES 

STIMULUS ORGANISM 

ONLINE ENVIRONMENTAL CUES INTERNAL STATES 

SHOPPING OUTCOMES 

** High Task Relevant Affect * Pre- 

Purchase Intentions 

* Navigational Characteristics * Attitude 

(Involvement in Purchase 

*Informativeness Decisions) 

* Effectiveness of Information Content Cognition 

* Structure * Exploratory Behavior * 

Purchase Intentions 

** Low Task Relevant 

* Entertainment of the Web site 

229 

* Involvement

230

Mapping Retailing Enterpreneur Decisions And Behavior 

Arch G. Woodside, Boston College 

Abstract 

Retailing entrepreneurial behavior requires timely modifications in the assumptions of 

entrepreneurs in response to environmental responses / non-responses to decisions / actions of 

the new enterprise. This article applies Axelrod's (1976) and Huff’s (1990; Jenkins & Huff 2002) 

approach to mapping strategic thought (causal mapping) to: (1) categorize how retailing 

entrepreneurs may respond to environmental feedback to their decisions and (2) assess the 

effectiveness of alternative implemented decisions in assisting organizational growth. A detailed 

example of causal mapping is presented for a retailing entrepreneurial case study; the example 

covers processes linking events, decisions, and activities in business start-up, growth, and failure 

of the enterprise. The article closes by suggesting that causal mapping is a valuable tool for 

advancing case study research. A research plan for future reports applying causal mapping in 

retailing entrepreneur studies is suggested. 

231

MAPPING RETAILING ENTREPRENEUR DECISIONS AND BEHAVIOR 

In the early 1970's Mintzberg (1973) asked an important question on strategymaking: 

how do organizations make important decisions and link them together to form strategies? 

Based on a review of the available literature he concluded that little systematic evidence exists 

about this important process, known in business as strategy-making and in government as 

policy-making. 

Fortunately, since the early 1970's substantial advances have occurred in theoretical and 

empirical research on how and why strategy-making occurs. One research stream on strategy- 

making has been labeled the "emergent stream" approach (Nutt 1993), whereby the researcher 

collects field data on many strategymaking case studies and immerses in the raw data describing 

each case to find the 

key decision making activities (e.g., Mintzberg, Raisinghani, and Theoret 1976; Woodside and 

Vyas 1987). The emergent stream approach contributes substantially to our understanding of the 

linkages of decision-makers perceptions of environmental signals, events, decisions, and 

activities through time. A criticism of the emergent stream approach: the mass of detail makes it 

difficult to analyze the data. As a result, researchers using the emergent stream approach are 

forced to examine a limited number of cases, making generalizability of their conclusions 

suspect. "Researchers using a small case data base may fail to discover important action-taking 

steps or fail to recognize the idiosyncratic nature of the steps that are discovered" (Nutt 1993, p. 

228). 

232

Applications of causal mapping are a second research stream on strategy-making. 

Causal mapping is a form of content analysis that isolates the key assertions within a 

document, such as a detailed case research report; the content analysis includes proposing 

linkages of causality among signals, events, decisions, and activities (e.g., Fahey and 

Narayanan 1989; Cossette 1992; Barr, Stimpert, &Huff 1992; Jenkins & Huff 2002). Causal 

mapping studies have produced many linkages of observations -> insights -> propositions, for 

example, such maps indicate that (1) both successful and unsuccessful firms quickly notice 

environment signals but (2) respond to these signals very differently (two research findings 

that also serve as insights into success and failure, and well as theoretical propositions for 

future research). Calm mapping results indicate that leaders in failing firms fail to change their 

beliefs, focus, and ways of behaving during courses of protracted downward spirals, while 

leaders in firms able to adapt to new environments do make such transformations (cf. Barr et 

al. 1992, p. 34); the implication of these results to entrepreneurial strategy-making may be 

that leaders in new ventures may need to transform themselves and adopt new modes of 

strategy making (adaptive or planning rather than entrepreneurial) when their enterprises 

passes out of the start-up phase into growth, maturity, and renewal phases. Back in 1973, this 

point was emphasized by Mintzberg for the planning mode of strategy-making, "Planning is 

not a panacea for the problems of strategy-making." Some situations require little or no 

planning, such situations favor adaptive strategy-making; others require limited planning, such 

situations favor entrepreneurial strategy-making. 

233

Attempts to use a framework to find patterns in decision making are a third research 

stream (cf. Nutt 1993). Both reliability and generalizability improve when patterns emerge that 

have the same activity steps proposed in a framework and the step sequences fit a large number 

of cases. "However, imposing a framework has the disadvantage of creating the appearance of 

orderliness in what may be a chaotic process. Also, imposing a framework may result in losing 

important messages that do not fit the framework" (Nutt, 1993, p. 228). Nutt (1993; 2003) 

backs-up his view (with field data) that the advantage of systematic description of decisions 

and using a large case data base outweigh the disadvantages of matching to an overly neat set 

of steps and the possibility of lost information. 

This article describes an exploratory application of causal mapping to retailing 

entrepreneurial research. We begin by reviewing a few propositions unique to entrepreneurial 

strategy-making from insight and the available literature on entrepreneurial studies; the 

propositions serve as a framework for linking many of the decisions and events that occur in 

entrepreneurial behavior. Next, a brief review of causal mapping is provided; causal mapping 

may be used to summarize proposed, deductively developed, frameworks, as well as to 

inductively analyze entrepreneurial behavior. Third, a generic causal model of entrepreneurial 

start-up, growth, and failure is described. Fourth, a detailed field case study of the model is 

examined. Finally, implications are provided for research on entrepreneur strategy-making and 

for improving marketing decisions by leaders of new enterprises. 

234

Propositions on Entrepreneurial Strategy-Making 

The four principal characteristics of the entrepreneurial mode of strategy making 

proposed by Mintzberg (1973) serve as propositions distinguishing this mode of strategy-making 

from others (e.g.. adaptive and planning modes). The following four propositions are based, in 

part, on Mintzberg's 1973 statements. (1) In the entrepreneurial mode, strategy-making is 

dominated by the active search for new opportunities; problems are secondary and receive less 

attention. (2) In the entrepreneurial organization, power is centralized in the hands of the chief 

executive who rules by fiat, not by committee or teams. (3) Strategy-making in the 

entrepreneurial mode is characterized by dramatic leaps forward in the face of uncertainty; the 

organization makes dramatic gains with high risk of failure. (4) Growth, not efficiency, is the 

dominant goal of the entrepreneurial organization; thus, the entrepreneur is focused on growth in 

sales volume, not cost control or profits. 

Stevenson, Roberts, and Grousbeck (1989) definition of entrepreneurship also may 

serve as a central proposition of entrepreneurial strategy-making: (5) the pursuit of opportunity 

without regard to resources currently controlled. An axiom of this proposition is that 

entrepreneurs search for specific opportunities, and the selection/formulation of new venture 

concept, to invest their energies into transforming into realities before searching for financial 

resources. Waiting to grow a nest-egg to invest and then searching for an investment 

opportunity is not representative of most entrepreneurial strategy-making. 

Results from Collins and Moore's (1970) study of 100 entrepreneurships suggest 

several additional propositions that distinguish entrepreneurial strategy making, for example, 

almost all successful entrepreneurs experience several failures before creating and 

implementing a successful enterprise. An axiom to this proposition is that several start-up 

235

attempts resulting in failure are necessary for learning effective deal-making. Collins and 

Moore (1970, p. 80) use a school-of-hardknocks analogy for describing this process: basic 

dealing is not a course for the faint hearted, or for the fair-weather entrepreneur. This course 

separates the men from the boys. There is no telling how many broken businesses and broken 

homes are left behind by students who enroll in the course, but who did not have the character 

to complete it. This is part of the high price paid for the education of entrepreneurs. 

For their most successful new enterprise, entrepreneurs find one or two sponsors, or 

mentors, to help in achieving success for a unique financial, or marketing, decision/action. 

Seeking and finding help from someone outside the firm who has both the expertise and ability 

to help is found almost always in successful case studies on entrepreneurships (Collins and 

Moore 1970, Chapter 5). 

Likely, the 1973 insights by Mintzberg offer important propositions that still need to be 

examined empirically about entrepreneurial behavior. To achieve long-term success and avoid 

mid-range failure (often between the third and sixth year of the new enterprise), the entrepreneur 

must dramatically alter his / her strategy-making from the entrepreneur mode to include the 

adaptive and planning modes. Axioms of this proposition include (a) flexibility in actions 

mustshift to become constrained, not bold; (b) almost linkages between events and decisions 

must be integrated, not loosely coupled; and (c) analytical, not judgmental, evaluations of 

proposals must be made. 

Thus, at some point during the growth stage of the successful new enterprise, the 

entrepreneur needs to recognize the need to change his/her mode of strategy making; most 

236

entrepreneurs fail to recognize this need and continue making bold moves, focus on maintaining 

growth (forecasting impossible sales volumes), and ignoring major problems. 

Causal Mapping 

Causal mapping is mostly used as a form of content analysis that isolates the key 

assertions within a document that deal with causality, existence, and/or categorization. The links 

among decisions, activities/events, and the interactions among persons can be summarized 

through time within causal maps. 

Causal mapping can be applied to descriptive data (i.e., thick descriptions reported in 

case studies) on the streams of decisions and behavior occurring through the stages of 

formulation and implementation of new enterprises. Coding schemes developed by Axelrod 

(1976) with modifications by Huff, Narapareddy, and Fletcher (1990) may be used to indicate 

causal and definitional relationships among the linkages in the causal maps; we used some of the 

coding categories developed by these researchers for the relationships proposed and found in 

entrepreneur strategy-making. The seven coding categories used here are summarized in Table 1. 

The symbol, "c," for choice criterion in making a decision is the one additional category not 

found in the codes used by Barr et al. (1993). 

Table 1 here. 

A "map" or series of related maps is then constructed by connecting specific events, 

decisions, and interactions of people with a symbol for the type of relationship observed. For 

example, consider the sentence, "entrepreneurs search for specific opportunities, and the 

selection/formulation of new venture concept, to invest their energies into transforming into 

realities before searching for financial resources." The linkages among activities decisions 

237

included in this sentence would be coded as follows: (search for specific opportunities) � 

(selection / formulation of new venture business model) � (search for financial resources). 

Three coders were trained to analyze several thick descriptions of cases on entrepreneur 

start-ups (in Fraser 1992). Following training and discussion on two of cases, the third set of 

codings were completed individually and tested to ascertain intercoder reliabilities. The average 

agreement among the three coders was a satisfactory 88 percent on the causal linkage codes 

assigned. Previous applications 

of near identical coding methods have achieved similar levels of high reliabilities (e.g., Axelrod 

1976; Barr, et al. 1992). 

Following training and the establishment of intercoder reliability, all three coders were 

assigned to code a fourth case description on entrepreneur behavior, the Millie Hand Cooked 

Potato Chip case (in Fraser 1992). The average intercoder reliability for the Millie case was 92.1 

percent. 

Working separately, the three judges were also asked to provide a parsimonious map 

that further summarized the key concepts and links in the Millie Case. The three judges then 

worked jointly to combine their three summary maps into the one described in the results 

section. 

After testing for intercoder reliability and developing the parsimonious, summary, map. 

The three judges also worked together to combine their three detailed maps of the stages found 

in entrepreneur behavior in the Millie case. In developing their individual maps the three 

judges had concluded that three stages had occurred in the Millie case: start-up, growth, and 

demise. 

238

Finally, the three judges worked as a team to evaluate 17 decision areas in each of the 

three stages. A seven-level evaluation system was used ranging from: 

• !!! for a remarkably effective decision/event 

• ??? for a disastrous decision/event. 

Because the evaluation assignments were made knowing the outcomes of the decisions and 

events, the assignments were made with few disagreements among the judges; disagreements 

were settled by discussion and consensus. Having the judges make evaluation assignments 

individually and testing for interceder reliability would be an improvement on this method. 

The Millie Case 

The "Millie's Hand Cooked Potato Chips" case is an undisguised thick description of 

entrepreneur behavior covering a five-year period, 1984 through 1988. Brian Shore was the 

entrepreneur who founded the company in 1984. Sales of the Millie hatched cooked chips 

(versus continuous cooked chips offered by larger competitors) increased from $500,000 to $3 

million during the brief life of the firm. 

Limitations. Causal mapping approaches of retailing entrepreneur case studies are 

limited in several ways. For research on entrepreneur behavior, additional work on inter coder 

reliability is needed to confirm the consistency of both creating concepts and coding linkages in 

such maps. Also, the accuracy of most published thick descriptions of real-life entrepreneur 

cases is questionable. The use of multiple research methods and the checking of facts by using 

multiple sources, and the re-checking of mental thoughts is not often described in such cases; 

thus, no evidence is provided in most of these cases on concurrent validity. 

239

Detailed causal maps are needed in large samples of entrepreneur case studies to test the 

propositions. The application presented here is exploratory only. Certainly causal mapping of 

entrepreneur behavior provides a research approach suitable for examining large numbers of 

such cases, similar to the research on formulation processes in organizational decision making 

reported by Nutt (1993). 

A Generic Causal-Map Model of One Category of Entrepreneur Behavior 

Figure 1 is the summary map developed by the three judges and based on their 

240

individual summary maps of the Millie case. Note that this summary map does reflect most of 

the defining propositions described as representative of entrepreneur Behavior. For example, 

the first box in the map verifies the sixth proposition that most entrepreneurs have a history of 

starting several new ventures that fail. 

Figure 1 here. 

In Figure 1, search and venture concept formulation is shown as leading steps before 

funding search. This sequence fits the fifth proposition and what Stevenson et al. (1989, p. 7) 

focus on to define "entrepreneurship as a behavioral phenomenon." The appearance of box 14, 

unrealistic sales growth forecasts, is a concept that may often appear in entrepreneur causal 

maps; this concept may be an outcome of the fourth proposition that growth, not efficiency, is 

the dominant goal of the entrepreneurial organization; thus, the entrepreneur is focused on 

growth in sales volume, not cost control or profits. The rapid increase in costs reported in box 13 

reflects the lack of cost control of the fifth proposition. 

Of course, empirically testing of the eight propositions is not possible based on a 

summary of one case. However, the summary presented in Figure 1 serves to illustrate that 

several propositions representative of entrepreneur behavior are likely to occur in causal maps of 

such behavior. Testing for the occurrence of each proposition is possible using large sample 

sizes of entrepreneur case studies (e.g., n > 100). To test the eighth proposition on transforming 

entrepreneur to planning modes of strategy-making would require case samples to include 

successes and failures that extend through long time periods, for example, five to ten years of life 

for each case. 

Detailed Causal Map of Entrepreneur Start-up, Growth, and Demise 

241

Start-up. Details of the stream of decisions, events and interactions between channel 

members that involved the Millie Company are summarized in Figure. 2. Note in Figure 2 that 

some signals are present (and not present) that support the third proposition discussed on 

entrepreneur behavior, that is, strategy-making in the entrepreneurial mode is characterized by 

dramatic leaps forward in the face of uncertainty; the organization makes dramatic gains with 

high risk of failure. For example, no detailed marketing evaluation or plan occurs between before 

or after box 4, when the new venture formulation and selection process is complete. No 

verification is made that some market target may actually prefer the new product versus 

competing products before the product and packaging is completed. Substantial funding is 

achieved in the face of high risk of failure. 


The decisions within Millie displayed in Figure 2 are made by Brian Shore; for the start- 

up power is centralized in the hands of the chief executive who rules by fiat, not by committee or 

teams. Thus, the second proposition is supported for the first stage in the case. Note that this 

finding holds even for the flavor of the new potato chip, Mr. Shore alters the product to suit his 

own taste (box 2 in Product Design); customers in the selected target market are not consulted-- 

no taste tests are done-before manufacturing. 

Marketing the product appears doomed for failure or limited success at the end of the 

start-up stage; no major supermarket chains agree to carry the new product because of lack of 

advertising to create customer awareness (box 13) and the failure of Mr. Shore to provide large 

retailers with expected payments for shelf space for new products (box 15). 

242

Growth. The growth stage of Millie Company is summarized in Figure 2. This causal 

map begins with a remarkably beneficial decision by Mr. Shore. He approaches a large 

Supermarket chain in the Ontario Province of Canada with a proposition to manufacture a 

Kosher potato chip for the Jewish holidays. This unique offer, marketing tactic, is accepted; for 

the first time, Millie is able to gain product placement in a large supermarket chain with stores in 

several Provinces (without offering shelf allowances). With the product's acceptance in one 

chain, the other' three largest chains quickly follow this lead (box 3). Thus, the third proposition 

described is supported: strategy-making in the entrepreneurial mode is characterized by dramatic 

leaps forward in the face of uncertainty; the organization makes dramatic gains with high risk of 

failure. 


. 

Note also that Mr. Shore continues to pursue rapid growth through-out the growth stage 

(see boxes 6 and 15 in Figure 3). He is unwilling to transform strategymaking from the 

entrepreneur mode to an adaptive, or planning, mode. 

Two decisions made by Mr. Shore to continue sales growth were judged to be very bad 

moves (??) by the three judges: the decisions shown in boxes 6 and 15. Millie's batch process 

versus competitors' continuous process requires substantially higher unit costs (see box 11 in 

Figure 2). 

Thus, offering the Millie product at a price lower than competitors' prices provides 

substantially less margin to cover costs and generate profits. Also, lessons from the Profit Impact 

of Marketing Strategies (PIMS 1980) include the folly of little market share brands companies 

attempting to compete head-on against much larger foes (box 15); successful small firms are 

243

more successful when they avoid bigger competitors and offer unique products to narrow market 

niches at higher prices 

(PIMS 1980). 

Demise. The final stage for the Millie company begins with a decision evaluated by the 

three judges to be a disastrous move: Mr. Shore purposively decreases quality by increasing the 

usage time of the frying oil in the manufacturing process (box 1 in Figure 4). The resulting 

product is substantially more greasy than prior manufacturing runs. Sales start to decline almost 

immediately (boxes 3 and 4 in Figure 4). 

Note also in Figure 4 that Mr. Shore never focuses on issues of lost sales but continues to 

focus his efforts on expanding market coverage: by moving more aggressively into the Ontario 

market (box 9) and attempting to hire a broker to add distributors (box 10). 


Evaluation of Outcomes of 17 Decisions Areas across the Three Stages. 

Table 2 includes the summaries of the three judges outcome evaluations of 17 decision 

areas. Note that between the start-up to the demise stages that a shift does occur in the number 

of decision areas be judged very good to very bad, for example, product design decisions shift 

from very good (box 6 in Figure 2) to remarkably good (box 1 in Figure 3) to very bad (box 15 

at the end of Figure 3) between the three stages. 

Table 2 here. 

Note also in Table 2 that the pricing strategy was judged very bad consistently across the 

three stages. Evidently, Mr. Shore did not consider that the selected target market of young, 

244

upwardly mobile, professionals, might be willing to pay a higher price for a better tasting with 

all "natural ingredients," Most likely, the high cost, low-price, strategy for the Millie chip helped 

the firm move to a negative cash flow and a lower quality image in customers' minds, than would 

have been achieved with the product's price being above competitors' prices. 

By the close of the company, only the outcomes associated with the packaging were 

considered positively. The outcomes of all other decision areas shifted negatively or remained 

negative. In Table 2, decision area outcomes indicated by both an exclamation point and 

question mark M) are to indicate that the outcome contains both positive and negative 

outcomes. For example, product quality is unique in the start-up and growth stages, but might 

have been made substantially better with some consumer taste tests of alternative recipes. 

Implications for Research on Retailing Entrepreneur Behavior 

Causal mapping may be used to both develop theoretical models of sets of propositions 

about processes representative of retailing entrepreneur strategy-making. Such deductively 

developed processes may be tested by comparing their predictions of the presence and sequence 

of decisions, events, and interactions of people with causal maps of actual entrepreneur behavior. 

Large size samples of thick descriptions of real-life entrepreneur cases may serve as data 

bases for developing empirical causal maps. Such large scale studies on decision processes are 

appearing in the literature of organization science. As such, causal mapping represents an 

advanced case study research method deserving wider use among researchers in entrepreneur 

behavior. 

References 

245

Axelrod, Robert (1976), "The Cognitive Mapping Approach to Decision Making," in Robert 

Axelrod, ed., Structure of Decision Making, Princeton, NJ: Princeton University Press, 

221-250. 

Barr, Pamela S., J. L. Stempert, and Anne S. Huff (1992), "Cognitive Change, Strategic Action, 

and Organizational Renewal," Strategic Management Journal, 13 (S), 15-36. 

Collins, Orvis, and David G. Moore (1970), The Organization Makers: A Behavioral 

Study of Independent Entrepreneurs, New York: Appleton-Century-Crofts. 

Cassette, Pierre, and Michel Audet (1992), "Mapping of an Idiosyncratic Schema," 

Journal of Management Studies, 29 (3), 325-347. 

Fahey, Liam, and V. K. Narayanan (1989), "Linking Changes in Revealed Causal Maps and 

Environmental Change: An Empirical Study," Journal of Management Studies, 26 (4), 

361-378. 

Follows, Scott B. (1992), "Millie s Hand Cooked Potato Chips," in Hanna Fraser, 

Entrepreneurship in Atlantic Canada, Acadia University, Wolfville, Nova Scotia: 

Atlantic Entrepreneurial Institute, 147-152. 

Huff, A.S., V. Narapareddy, and K.E.Fletcher(1990), "Coding the Association of Concepts," in 

A.S. Huff, ed., Mapping Strategic Thought, Chichester, U.K.: Wiley, 311-326. 

Huff, A. S. (1990), "Mapping Strategic Thought," in A. S. Huff, ed., Mapping Strategic 

Thought, Chichester, U.K.: Wiley, 88-115. 

Jenkins, Mark, and A. S. Huff (2002), Mapping Strategic Knowledge, Thousand Oaks, CA: 

Sage. 

Mintzberg, Henry (1973), "Strategy-Making in Three Modes," California Management Review, 

16 (2), 44-53. 

246

Mintzberg, Henry, D. Raisinghani, and A. Theoret (1976), "The Structure of ‘Unstructured’ 

Decision Processes,” Administrative Science Quarterly, 21, 246-275. 

Nutt, Paul C. (1993), "The Formulation Processes and Tactics Used in Organizational Decision 

Making," Organization Science, 4 (2), 226-252. 

PIMS, The PIMS Program (1980), Cambridge, MA: The Strategic Planning Institute. Stevenson, 

Howard H., Michael J. Roberts, and H. Irving Grousbeck (1989), NewBusiness Ventures 

and the Entrepreneur, Homewood, IL: Irwin. 

Woodside, Arch G., and Niran Vyas (1987), Industrial Purchasing Strategies, Lexington, MA: 

Lexington Books. 

247

1. Brian Shore starts several 

+ 

2. B.S. searches for new 

7 Search for financing 

of new venture 

E’s that result 1 E in starts failure several new venture 2 opportunity Search for new 

ventures that fail 

venture opportunities + 

c 

+ 

+ 

3. Products generates services 4 Snack food, potato 

even in hard times 

e 

chips chosen 

e 

m + 

6 Final product design 4 Market + analysis mfg. 

5 Canadian retail sales in 

analysis 

1984 are $500 million; 3 

6 B.S. goal is 2% of 

+ 

dominating 

competitors 

industry sales 

m 

+ 

+ 

7 Target market 

+ 9 Manufacturing 

selection and marketing 

process finalized and 

plan 

runs completed 

+ 

e 

+ 

m 

3 Venture concept 

e 

formulation 

8 $30.000 from 

silent partner 

e + m 

9 $64.000 from 

5 sixth Funding bank search 

and approached. success 

10 $10,000 from 

federal 10 Distribution government 

small problems business occur; loan 

secondary system 

developed 

Product Design 

7 Only firm with m 

8 Package 

hand cooked chip + design a 

e m 2 B.S. alters to suit his 

1 75 year-old 

in Canada 

own tastes 

Mennonite 8 recipe Pricing, Product, Promotion, 

12 Major, favorable, + e 11 Unique 

and Place Destination levels 

distribution development 

e 

marketing action 

occurs 

specified e 

6 Thicker, more nutritious = 9 “Millie’s sound taken by CEO 

3 Batch Process 

and flavorful chip 

liked by B.S. 

e 

+ 

+ 

e 

e 

5 No preservatives, no m.g.; 

sunflower 

+ 

oil; chips 

10 “Natural ingredients” 15 Unrealistic and sales 

statement on growth mfg process forecasts 

4 Chips hand-stirred; cholesterol free 14 Attempt to reduce 

cooking time set by sight 

13 Rapid growth in 

+ 

+ 

costs by decreasing 11 Cost 20% higher than 

product quality conventional 17 chips Customer 

sales and costs; 

+ franchise declines; 

widespread distribution + 

rapid snowball 

acceptance 

16 Pricing strategy: 

decline in sales 

Implementation 

+ 

be lower than larger 

competitors 

13 No advertising; lots a 14 Distribution: independent grocery 

12 Target 

of free samples 

and convenience stores 

market: Yuppies 

Key: E= entrepreneur a 

+ 

+ 

15 B.S. refuses to provide shelf 

16 Chains refuse to 17 Unique Go to 

allowances to get listings in chains carry new 

FIGURE 

product 

1 

product stage 2 

Generalized Cause Map of Entrepreneur Start-Up, Growth, and Failure Process 

FIGURE 2 a 

Cause Map of Failure Process of Entrepreneurship: 

Millie’s Handcooked Potato Chips; Stage 1, Start Up 

a key: E= entrepreneurship; m.g.= monosodium glutamate 

+ 

248 

+ 

- 

+

1 B.S. approaches 

Steinberg chain with 

offer to provide Kosher 

chip 

2 Unique product 

design matched to value 

of narrow target market 

9 New $200,000 foam- 

filled 

packaging 

machine 

purchased; 

variety increased 

13 B.S. forecasts 

sales of $13 million 

next year 

14 Production 

operation moves into 

$1.3 million 

plant 

financed by $250, 000 

loan from SBDC 

e 

+ 

+ 

m 

a 

m 

+ 

+ 

3 Millie’s accepted in 

3 large chains; Sobey’s, 

The Food Group, IGA 

7 B.S. contracts with 

independent wholesalers 

through Maritime 

Provinces 

m 

8 Wholesalers not 

visited by B.S.; no store 

visits 

249 

+ 

+ + 

11 Stores sold out of 

Millie’s in 2 weeks 

+ 

15 B.S. wants to 

compute in 

conventional chip 

market and mfg. thin 

chip, “Archie’s” 

FIGURE 3 

Cause Map, Stage 2, Growth 

+ 

+ 

m 

a 

+ 

4 Distribution 

expands to PEI and 

Newfoundland 

+ 

+ 

5 B.S. goal is to 

increase sales 

volumes 

+ 

6 Millie positioned 

as low-price 

competitor 

10 1987, first large 

national order received 

12 First 

international order 

received from 

Taiwan importer 

16 Price strategy 

lower than national 

large, competitors 

products 1987 sales 

= $3 million 

- 

-

+ 

1. 

B.S. 

‘Maximizes use 

of frying oil to 

reduce 

8. 

Some 

wholesalers 

drop product 

due to falling 

sales 

m 

9. 

B.S. continues to 

look for 

expansion in 

Ontario 

m 

2. 

No testing done 

on oil quality 

+ 

+ 

13. 

Sales drop further 

a 

+ 

7. 

Millie’s sales 

drop 

dramatically 

a 

3. 

Millie’s chips 

became 

interestingly 

greasy 

250 

6. 

Government 

introduces new 

sales tax on snack 

food 

10. 

11. 

B.S. meets 

+ 

Miles several 

+ Tentative 

m 

agreement 

times to act as 

Ontario broker reached: Miles to 

be broker 

m 

+ 

+ 

14. 

Touche Ross appointed to be 

receiver of Millie’s: Millie’s 

bankrupt 

Figure 4 

Cause Map, Stage 3, Demise 

+ 

4. 

Sales start to 

decline in June 

1987 

5. 

Price war starts 

among 3 large 

national 

competitors 

12. 

B.S. stops 

returning 

Miles’ 

phone calls 

m

An Empirical Examination of Retail Convenience for In-Store and Online Shoppers 


Michelle Bednarz, Mississippi State University 

Abstract: 

In this study, we develop and assess multidimensional scales of retail convenience for in-store 

and online shoppers. Principal components analysis and confirmatory factor analysis reveal five 

dimensions of retail convenience for in-store shoppers, and four dimensions of retail 

convenience for online shoppers. Analysis of variance results indicate that online shoppers 

benefit more from access convenience and search convenience than in-store shoppers. 

Implications for retailers and areas for future research are also addressed. 

Introduction 

The concept of convenience has been an important topic in marketing for a time period 

spanning approximately 80 years (Copeland 1923; Gardner 1945; Kelley 1958; Brown 1989, 

1990; Seiders, Berry, and Gresham 2000; Berry, Seiders, and Grewal 2002). However, much of 

this research focuses on convenience goods, rather than convenience as a benefit realized by the 

consumer. Because today’s consumer is more time-starved than ever, it is appropriate to more 

carefully consider convenience as a concept of utmost importance. 

Foote (1963) characterized the consumer of the future (in the year 2000) as one whose 

primary constraints are no longer money, but time and learning. His prediction is correct. 

Today’s consumers consider the resource of time as valuable as money (if not more valuable). 

Bhatnagar, Misra, and Rao (2000, p.98) confirm this by stating that “the consumer maximizes 

his utility subject to not only income constrains but also time constraints.” 

As time becomes a more valuable consumer resource, convenience also provides added 

value. Michael J. Sansolo, senior vice-president at the Food Marketing Institute, illustrates the 

251

degree of importance consumers place on convenience by stating, “Give shoppers a choice 

between lower prices or more convenience and convenience will win every time” (Forster 2002, 

p. 120). Online shopping is particularly well suited to offer convenience to time-starved 

consumers (Kaufman-Scarborough and Lindquist 2002). In an otherwise lackluster economy, 

the strong growth rates of companies such as eBay, Land’s End, and Amazon have occurred 

because they provide customers with what they want – its all about convenience (Green 2002). 

When examining the literature related to the convenience construct, many gaps exist, 

particularly in a retail context. While it is assumed that consumers often seek convenience in 

shopping situations, we know little about the individual dimensions of retail convenience. 

Seiders, Berry, and Gresham (2000) propose four dimensions of convenience that are 

particularly relevant for retailers: access, search, possession, and transaction convenience (see 

Table 1). An empirical investigation of how these convenience dimensions provide a benefit to 

consumers is warranted. Specifically, do in-store shoppers and online shoppers seek and obtain 

different types of convenience? This study considers that different structures of retail 

convenience may exist, depending on the shopping format chosen by the consumer. The 

argument developed here is that different shopping formats deliver different types of 

convenience, and a consumer’s preference for a specific type of convenience determines his or 

her decision to choose a particular shopping format – either online or at a traditional store. 

This study is important for several different reasons. First, it follows the logical 

progression of research in this area. Retailers may now better understand the existence of 

different convenience dimensions, but how can they use this new knowledge to better satisfy 

customers? As Berry (2001, p. 136) states, “Superior retailers understand that people’s most 

precious commodity in the modern world is time and do everything they can to save as much of 

it as possible for their customers.” By developing a better understanding of the different 

252

dimensions of retail convenience and by understanding which dimensions of convenience are 

most important to customers, retailers will be able to better understand how to meet customers’ 

needs, thus improving customer satisfaction. If properly implemented, retail convenience could 

become a source of sustainable competitive advantage for retailers. 

This paper is organized in the following manner: A review of the literature relevant to this 

study is presented. Scale development procedures are followed in order to more carefully 

examine the specific dimensions of retail convenience for both in-store and online shoppers. A 

different model for each type of shopping format is hypothesized and tested using structural 

equation modeling techniques. Results are given, followed by a discussion that addresses the 

implications of the findings for retailers with traditional stores and/or an online presence. 

Limitations and directions for future research are also provided. 


The concept of convenience first appeared in the marketing literature with Copeland’s 

(1923) classification of goods. Copeland suggests that by classifying goods according to his 

tripartite structure (convenience, shopping, or specialty goods), marketers can determine the type 

of store in which the product should appear and can determine the appropriate concentration of 

distribution. Convenience goods are those lower-priced goods which consumers are familiar 

with and which are purchased from easily accessible outlets. Gardner (1945, p. 275) provides a 

further description: 

Convenience goods are articles of daily purchase…which are insignificant in 

value or are needed for immediate use. These goods are, to a considerable extent, 

bought at the most convenient place without a comparison of values…. 

As marketers continued to develop other product classification systems, convenience 

goods remained an essential staple, re-appearing in several other schemata (Bucklin 1963; Kaish 

1967; Holbrook and Howard 1977; Enis and Roering 1980; Murphy and Enis 1986). Thus, the 

253

initial use of the word “convenience” in the marketing literature was as an adjective describing a 

class of consumer goods. Many researchers have conducted studies covering the vast domain of 

convenience goods, including convenience foods (e.g., frozen dinners, ready-to-eat cold cereal, 

fast-food restaurants) (Crist 1960; Anderson 1972; Reilly 1982; Darian and Cohen 1995), 

convenience time-saving durables (e.g., dishwashers, microwave ovens, washers and dryers) 

(Anderson 1972; Reilly 1982), and time-saving services (e.g., child care, house cleaning 

services, lawn care services) (Brown 1990). 

Over time, the use of the word “convenience” has changed/evolved from a descriptor of 

products into its own unique concept. The more complete definitions of convenience contain 

one common element - they all view convenience as any attribute that reduces the non-monetary 

costs associated with a product (Kelley 1958; Kotler and Zaltman 1971; Etgar 1978). For 

purposes of this study, retail convenience (at its most general level of abstraction) is defined as 

the consumer’s perceptions of time and effort costs associated with making a purchase. As the 

reduction of time and effort costs is a motivating factor for consumers to seek convenience, these 

concepts are described more fully below. 

Time and Effort Literature 

Time is a limited, scarce, and therefore valuable consumer resource (Jacoby, Szybillo, and 

Berning 1976). As a limited, scarce, and valuable resource, consumers are always looking for 

ways to save time, be it through convenience products, convenience services, convenience 

durables, or more recently through convenience shopping. One reason why saving time has 

become so important to consumers is because time, unlike money or other resources, is fixed and 

therefore incapable of being expanded (Berry 1979; Gross 1987). According to Berry (2001, 

p.136), “Superior retailers understand that people’s most precious commodity in the modern 

world is time and do everything they can to save as much of it as possible for their customers.” 

254

Like time, effort is also a valuable consumer resource which is characterized by 

consumers’ energy expenditures. Activities requiring great amounts of energy on the 

consumers’ part are considered inconvenient, while activities requiring minimal energy on the 

consumers’ part are considered convenient. As a completely different type of non-monetary 

cost, consumers’ effort expenditures not only influence perceptions of convenience (Seiders, 

Berry, and Gresham 2000), but also influence levels of satisfaction (Lovelock 1983). 

Human and cognitive effort has been the topic of many studies appearing in other 

disciplines including operations management, psychology, decision theory, and economics 

(Youngdahl and Kellog 1997; Bettman, Johnson, and Payne 1990). However, consumer effort 

expenditures have received significantly less attention within the marketing literature (Berry, 

Seiders, and Grewal 2002), perhaps because product or service attributes that aim to save effort 

are perceived as being the same as, and as a result are lumped together with, those that aim to 

save time (Brown 1990). Thus, in a retail context, consumers view both time and effort as costs 

that are associated with obtaining the desired product and/or service. In the decision-making 

process, consumers must decide if the costs of obtaining the desired product/service are worth 

the benefits they will receive from the purchase. 

Cost-Benefit Analysis 

The driving force that motivates consumers to seek convenience involves saving both time 

and effort expenditures. According to Prest and Turvey (1965, p. 683), “Cost-benefit analysis is 

a practical way of assessing the desirability of projects” and “implies the enumeration and 

evaluation of all the relevant costs and benefits.” When deciding between several alternatives 

(e.g., whether to shop in a traditional store or in a virtual store), consumers determine the costs 

and benefits associated with each and compare those costs to the benefits. 

255

In the decision-making process, time and effort are significant consumer costs and are 

often considered along with money and other resources (Jacoby, Szybillo, and Berning 1976; 

Bhatnagar, Misra, and Rao 2000). Consumers choose online shopping because it facilitates the 

steps in the decision process, saving consumers both time and effort in the search, alternative 

evaluation, and purchase stages. Bhatnagar, Misra, and Rao (2000) apply cost-benefit analysis 

to explain why consumers choose to shop online: 

[Internet stores] tend to reduce the time the consumer spends on shopping (travel 

time, time spent parking, time spent traveling from the parking lot to the store, 

time spent in the checkout lines) either directly or indirectly. … The only time 

component remaining is the time spent browsing the Web sites (which 

corresponds to the time spent browsing the aisles in the more traditional sources 

of retailing). Therefore, a great attraction of the Internet is the convenience that it 

affords (p. 98). 

In other words, the benefits of shopping online (easy access to retailers) outweigh the costs (not 

obtaining the product immediately). Depending on what type of convenience is most important 

to the consumer, the cost-benefit analysis could lead to a different shopping format decision. 

The different types of convenience that consumers may choose to seek are described below. 

The Multidimensional Nature of the Convenience Construct 

Table 1 presents the various definitions and dimensions of convenience proposed by 

researchers in this area. Brown (1989) is the first to define the construct of convenience, and he 

focuses on the need for a definition that reflects the term’s multidimensional nature. The 

definition of convenience proposed by Brown (1989, 1990) contains five different dimensions: 

time, place, acquisition, use, and execution. Note that all five of the dimensions reflect the idea 

of saving consumers’ time and effort expenditures. 

-------------------------------- 


-------------------------------- 

256

Berry, Seiders, and Grewal (2002) identify and define five dimensions of convenience 

applicable to the services arena: access, decision, transaction, benefit, and post-benefit. Seiders, 

Voss, Grewal, and Godfrey (2003) empirically investigate these dimensions of service 

convenience and their relationship to value perceptions and behavioral intentions. Their findings 

suggest that decision convenience has a positive effect on value perceptions; access convenience 

has no significant effect on either value perceptions or behavioral intentions; transaction 

convenience has a positive effect on behavioral intensions; and benefit convenience has a strong 

positive effect on both value perceptions and behavioral intentions. 

Seiders, Berry, and Gresham (2000) identify and define four dimensions of convenience 

that are specific to retailers. These four dimensions of retail convenience are essential for the 

purposes of this study. They are discussed in greater detail below. 

Access convenience. Access convenience is defined as “the speed and ease with which 

consumers can reach a retailer” (Seiders, Berry, and Gresham 2000, p. 81). This access may 

occur in person, over the phone, through a computer, or in other ways. Access convenience is an 

extremely important dimension of retail convenience, because if the consumer cannot reach the 

retailer, then all other dimensions of retail convenience are meaningless. In other words, if the 

consumer cannot reach the retailer, then they will never be given the opportunity on that 

particular shopping attempt to make a decision, to complete a transaction, or to possess the 

desired product from the retailer. 

Consumer decision-making is significantly influenced by both the speed and ease with 

which consumers can make contact with retail outlets. Traditional retailers may improve access 

convenience by operating from a location that is easy to get to, near to most consumers, and near 

to other frequently visited stores (Seiders, Berry, and Gresham 2000). Online retailers are 

particularly suited for access convenience, as it provides the opportunity for consumers to shop 

257

at home 24 hr/7 days a week (Hofacker 2001). Morganosky and Cude (2000) found that the 

main reason consumers choose to purchase groceries online is that it eliminates travel time to 

and from the store. 

Although it is an important aspect of retail convenience, providing access convenience 

alone will not necessarily lead to sales. A virtual or physical store may be easy to access, but at 

the same time slow or difficult to use (Seiders, Berry, and Gresham 2000). In order to facilitate 

the decision-making process, the retailer must also provide the information necessary for the 

consumer to make the best purchase decision. 

Search convenience. After access convenience reduces the time and effort necessary to 

reach a retailer, search convenience then eases consumers through the shopping process by 

helping them make their purchase decision. Search convenience is “the speed and ease with 

which consumers identify and select products they wish to buy” (Seiders, Berry, and Gresham 

2000, p. 83), and includes effective interactive customer systems, store design and layout, 

product displays, store signage, and knowledgeable salespeople. 

Search convenience is extremely important for retailers operating virtual stores. One of 

the greatest consumer benefits of shopping online is the advanced search capability and the ease 

with which consumers can locate product information (Hof 2001; Shop.org 2001; Burke 2002). 

Wolfinbarger and Gilly (2001, p. 35) support this by stating that the “the online medium 

facilitates utilitarian behavior as search costs for product information are dramatically reduced.” 

Thus, shopping on the Internet facilitates search convenience. 

Possession convenience. Seiders, Berry, and Gresham (2000, p. 85) define possession 

convenience as “the speed and ease with which consumers can obtain desired products.” 

Included within the domain of possession convenience are in-stock merchandise, timely 

production, and timely delivery of merchandise. One of the main reasons why consumers 

258

choose traditional stores over virtual stores is the ability to actually leave the store with the 

desired product. Online shoppers must wait for their order to be processed and then wait for 

delivery to their home or office. For consumers shopping online, low possession convenience 

may be seen as a cost rather than a benefit. 

Transaction convenience. Transaction convenience is defined as “the speed and ease 

with which consumers can effect or amend transactions” (Seiders, Berry, and Gresham 2000, p. 

86) and includes the checkout process as well as the return process. Hence, transaction 

convenience addresses the ease of doing business with a retailer during and/or after the act of 

purchase. One benefit of online shopping is that consumers never have to wait in line at virtual 

stores, thus enhancing transaction convenience (Wolfinbarger and Gilly 2001). Traditional 

stores as well as virtual stores with quick checkouts and easy return policies also rank high in 

transaction convenience. 

The four dimensions of retail convenience share a common element - they each save the 

consumer time and effort in a unique way. Whether shopping online or in a physical store, 

consumers seek these various convenience dimensions to reduce time and effort costs associated 

with consumer decision-making. The next section addresses the different models of retail 

convenience for each shopping format. 

A Model of Retail Convenience for In-Store and Online Shoppers 

Consumers often cite convenience as a major benefit of shopping online (Szymanski and 

Hise 2000; Hoffman 2000; Childers, Carr, Peck, and Carson 2001; Wolfinbarger and Gilly 

2001). We argue that a consumer’s preference for a specific type of convenience influences his 

or her preference for a particular shopping format - either online or at a traditional store. Figure 

1 contains the hypothesized models to be tested in this study. 

-------------------------------- 

259


-------------------------------- 

As the dimensions of convenience have only begun to be examined empirically, the first 

hypothesis relates to the different structures of retail convenience for in-store and online 

shoppers. We concur with Seiders, Berry, and Gresham (2000) that there exists four distinct 

dimensions of retail convenience. For traditional in-store shoppers, we feel an additional 

dimension of convenience should be considered. Hoffman (2000) found that in-store shoppers 

consider it convenient to ask for and receive assistance from a salesperson in making a purchase, 

especially if the consumer lacks the necessary information to complete the purchase himself. 

Particularly for purchases that require extensive problem solving, if the consumer lacks the 

necessary information, he will seek that information from expert problem solvers like 

knowledgeable salespeople (Dunn, Thomas, and Lubawski 1981; Hoffman 2000). Thus, 

“assisted search,” defined as the speed and ease with which consumers identify and select 

products they wish to buy due to assistance from a salesperson, is an additional convenience 

dimension applicable to traditional retailers. It is not appropriate, however, to consider search 

features on a retailer’s website as assistance in the same manner in which it takes place at a 

physical store. It is the knowledge, training, and advice from a human salesperson that may 

provide convenience in a way that a virtual store cannot. Thus, the following is hypothesized: 

H1a: For traditional in-store shoppers, retail convenience is a multidimensional 

construct containing five distinct dimensions: access, search, assisted search, 

transaction, and possession. 

H1b: For online shoppers, retail convenience is a multidimensional construct 

containing four distinct dimensions: access, search, transaction, and possession. 

In addition to testing the multidimensional nature of retail convenience, we also 

hypothesize about the relationship of these dimensions to in-store and online shoppers. As 

previously stated, the dimension of access convenience is defined as “the speed and ease with 

260

which consumers can reach a retailer” (Seiders, Berry, and Gresham 2000, p. 81). Consumers 

who frequently shop at online retail stores do so because they can shop from the comfort of their 

homes and at any time of the day or night. The ability to reach the retailer at a time most 

convenient to the consumer (access convenience) is certainly a benefit of online shopping. 

Compared to shopping at traditional locations, shopping online saves the consumer travel time to 

the location, time spent parking, and time spent traveling from the parking lot to the store 

(Bhatnagar, Misra, and Rao 2000). With traditional locations, however, customers are required 

to adjust their preferred shopping time to fit within the retailer’s hours of operation. 

The concept of cost-benefit analysis (Prest and Turvey 1965) may be used to explain why 

consumers who shop online benefit most from access convenience. Online shoppers believe that 

the benefit of time saved by being able to access retail outlets from the comfort of their home or 

office at anytime of the day or night far outweighs the costs of delayed merchandise possession 

and the risks associated with shopping online (Wolfinbarger and Gilly 2001; Morganosky and 

Cude 2000). In-store shoppers receive the benefit of immediate possession of their purchases at 

the sacrifice of convenient store access. Formally stated: 

H2: Online shoppers realize the benefit of access convenience more than 

traditional in-store shoppers do. 

Consumers who value search convenience do so because of reduced time spent looking for 

a particular product and finalizing their product decision. Online shoppers are more likely to 

realize the benefit of search convenience than in-store shoppers, simply because they are able to 

compare product and pricing information without having to leave home. Benefits falling within 

the dimension of search convenience include site design (Szymanski and Hise 2000), navigation 

(Childers, Carr, Peck, and Carson 2001), and selection and availability of product information 

(Wolfinbarger and Gilly 2001). Many consumers who choose to shop online do so because of 

261

the organization and design of the website and the ease of navigation. E-tailers having a user- 

friendly site design facilitate search convenience as consumers arriving at such a site can easily 

find exactly what they are looking for. Also considered in the dimension of search convenience 

is the product selection offered by the retailer. A virtual retailer is not limited by shelf space like 

a traditional retailer; therefore, they are usually able to offer a wider selection of products. The 

ease and speed with which the consumer can find exactly what he or she wants is key. The final 

benefit classified under search convenience is the availability of product information. How 

much information is available directly from the e-tailer’s virtual store? Do they supply 

comparisons to similar products offered by competing e-tailers? This is also important as many 

consumers turn to the Internet specifically for their information search. To conduct as extensive 

a search via traditional retailers would consume considerable amounts of time and effort. This 

leads to hypothesis 3: 

H3: Online shoppers realize the benefit of search convenience more than 

traditional in-store shoppers do. 

Customers seeking possession convenience prefer to save the time associated with actually 

obtaining the desired product (Seiders, Berry, and Gresham 2000). These consumers would 

rather spend time on other aspects of their shopping experience (e.g., getting to the store’s 

location and physically moving through the store to find exactly what they want) in exchange for 

not having to wait on parcel delivery to have their desired purchase in their possession. 

Applying cost-benefit analysis to this situation (Prest and Turvey 1965), consumers valuing 

possession convenience prefer to shop at traditional brick-and-mortar stores because the benefit 

of having the desired product in their hands at the end of the shopping trip outweighs the costs 

associated with traveling to the physical location and physically searching through the store’s 

shelves to find exactly what they want. In other words, they do not mind putting in extra 

262

shopping time as long as the desired product is in their possession immediately as a result of the 

shopping trip. Hypothesis 4 states the following: 

Construct Measurement 

H4: Traditional in-store shoppers realize the benefit of possession convenience 

more than online shoppers do. 

Research Method 

In order to measure the different dimensions of convenience, appropriate scale 

development procedures were followed (Churchill 1979; DeVellis 1991; Spector 1992). An 

initial survey containing several open-ended questions was administered to 196 students enrolled 

in upper-level marketing courses at a major Southeastern university. Questions such as “Please 

describe below what the word ‘convenience’ means to you” and “Describe as specifically as 

possible what your ideal convenient shopping experience would be like” were asked in order to 

develop the most appropriate phrases to capture each dimension of retail convenience. Many 

items were developed for each dimension, then 14 expert judges were asked to select the items 

that most closely matched the definitions of each dimension. A pretest was then conducted in 

order to improve/clarify question wording and instructions. These pretest surveys were given to 

a sample of seventy-five upper-level undergraduate marketing students to determine the 

necessary modifications. This process resulted in slightly different item wordings for in-store 

and online shoppers. Table 2 contains the final set of items for each convenience dimension. 

Sampling Procedure 

-------------------------------- 


-------------------------------- 

The final survey version was administered to a convenience sample consisting of both 

students and non-students. Marketing students enrolled in upper-level undergraduate consumer 

263

ehavior courses participated as both respondents and recruiters. Each student was asked to 

complete the survey and to recruit one other non-student to also complete the survey. Non- 

student names and phone numbers were collected and randomly checked to ensure authenticity. 

This process resulted in 308 total usable surveys (45% male, mean age 30); 241 completed their 

last major purchase in a traditional store, while 67 completed their last major purchase online. 

When answering the questions related to retail convenience, respondents were asked to 

think about their most recent major purchase. This was done to increase the chance that 

respondents would remember their purchase in as much detail as possible. Those respondents 

who made their most recent major purchase in a traditional store were directed to the scale items 

for in-store shopping, while those respondents who completed their most recent major purchase 

online were directed to the “online shopping” section of the survey. 

Statistical Technique 

Several statistical techniques were used to analyze the data. In order to test Hypothesis 1, 

statistics typically used in scale development were employed, including Cronbach’s alpha, 

principal components analysis, and confirmatory factor analysis using LISREL 8.3 for Windows. 

In order to test Hypotheses 2 through 4, analysis of variance was used to compare scale means of 

access, search, and possession convenience for in-store and online shoppers. 

Results 

In order to test Hypothesis 1, statistical procedures commonly used in scale development 

were employed. Reliabilities for each dimension of retail convenience were calculated using 

Cronbach’s alpha. Principal components analysis with varimax rotation was also undertaken in 

order to determine if five distinct dimensions were present for in-store shoppers, and four 

distinct dimensions were present for online shoppers. A more stringent confirmatory factor 

264

analysis using LISREL 8.3 for Windows was also undertaken in order to assess convergent and 

discriminant validity. Results for in-store shoppers are presented in Table 3, while results for 

online shoppers are presented in Table 4. Correlation matrices, means, and standard deviations 

of the scale items are provided for in-store shoppers in Table 5 and online shoppers in Table 6. 

-------------------------------- 


-------------------------------- 

-------------------------------- 


-------------------------------- 

-------------------------------- 


-------------------------------- 

-------------------------------- 


-------------------------------- 

Since all of the items within each dimension of retail convenience are reflective of their 

appropriate definitions, Cronbach’s alpha was used in order to examine the reliability of each 

dimension. As can be seen in the tables, these reliabilities were quite high, ranging from 0.7649 

to 0.9722. The dimensions of online retail convenience performed particularly well, as all were 

above the 0.9 level. 

For in-store shoppers, five dimensions of retail convenience were hypothesized to exist, 

while four dimensions of online retail convenience were hypothesized. In order to more 

carefully examine the dimensionality of these constructs, principal components analysis with 

varimax rotation was undertaken for each group separately. For in-store shoppers, five 

components with eigenvalues greater than one were extracted from the data. Together, these five 

components explain 77.9% of the total variance. For online shoppers, four components with 

eigenvalues greater than one were extracted; together, they explain 84.4% of the total variance. 

Thus, there is strong evidence in support of Hypothesis 1; the dimensionality of retail 

265

convenience is different for the two different types of shopping formats. Evidence of 

discriminant validity is provided by the fact that these items did not want to load on other 

components with which they were not supposed to be associated. Further, evidence of 

convergent validity is provided by the high loadings of each item on their respective component. 

A more rigorous examination of the validity of these scales was conducted with 

confirmatory factor analysis using LISREL 8.3 for Windows. As can be seen in Tables 3 and 4, 

the statistically significant λ x parameter estimates provide evidence of convergent validity. 

Additionally, the majority of the squared multiple correlations (SMCs), defined as the percentage 

of variance in each item that is explained by the latent construct of interest, are above 50%, 

indicating that each item performed well in capturing its construct of interest. 

Modification indices associated with λ x and Θ δ reveal some problems with discriminant 

validity. Specifically, there were seven modification indices in the λ x matrix for in-store 

shoppers, indicating that items search2, help3, possess1, and possess2 wanted to be associated 

with other dimensions of convenience. Additionally, there were 13 modification indices in Θ δ 

that indicate some error terms wanted to correlate. These changes were not incorporated in the 

model, as theoretically it does not make sense to do so. These problematic modification indices 

explain the poor overall fit: chi-square=498.08, 198 df, p=0.00; RMSEA=0.08; RMR=0.06; 

GFI=0.85. 

Problematic modification indices also surfaced in the measurement model for online 

shoppers. Three items (access2, possess2, and possess4) wanted to be associated with other 

constructs, while 6 error terms wanted to correlate. Again, these changes were not implemented, 

resulting in poor overall model fit: chi-square=206.34, 97 df, p=0.00; RMSEA=0.13; 

RMR=0.09; GFI=0.71. The overall fit does substantially improve if paths are freed as suggested 

by the modification indices. Therefore, confirmatory factor analysis results are mixed; there is 

266

strong evidence of reliability and convergent validity, but discriminant validity must be 

improved in further scale refinement attempts. 

For Hypotheses 2 through 4, analysis of variance (ANOVA) was performed in order to 

compare the mean scale scores for online and in-store shoppers. Hypothesis 2 states that online 

shoppers benefit more from access convenience than in-store shoppers. A comparison of scale 

means for these two groups yields an F statistic of 37.59 (p=0.000). Access convenience is more 

beneficial for online shoppers (scale mean=6.55) than in-store shoppers (scale mean=5.65), 

providing evidence of support for H2. Hypothesis 3 states that online shoppers benefit more 

from search convenience than do in-store shoppers; ANOVA here produces an F statistic of 8.57 

(p=0.004). The scale mean of 5.88 for in-store shoppers is significantly less than the scale mean 

of 6.26 for online shoppers; thus, there is evidence of support for H3. Finally, hypothesis 4 

predicts that in-store shoppers benefit from possession convenience more than online shoppers. 

The F statistic from this ANOVA is significant (F=11.07, p=0.001); however, the scale means 

indicate that the reverse is true. The possession scale mean for in-store shoppers is 5.60, while 

the scale mean for online shoppers is 6.23. This finding is quite counter-intuitive; possible 

reasons for this result are provided in the next section. 

Discussion 

This study provides an important contribution to the marketing convenience literature 

because it is the first of its kind to utilize the different dimensions of retail convenience to better 

understand customer needs. The support for Hypothesis 1 suggests that the dimensions of 

convenience are different for the two different types of shopping formats. The expansion of 

Seiders, Berry, and Gresham’s (2000) four dimensions of retail convenience to include “assisted 

search” is warranted. The five-component solution resulting from the principal components 

analysis is evidence of the existence of this newly proposed dimension. Further work in this area 

267

should seek to establish a similar convenience dimension for online shoppers. Technological 

capabilities that allow customers to experience “live help” or to receive recommendations based 

on what he has in his shopping bag may be considered a different type of “assisted help” than 

what is provided in a traditional store. It may be difficult, however, to obtain a large enough 

sample of online shoppers who regularly use these convenience-oriented features. 

The scale items crafted to measure the different dimensions of retail convenience possess 

high reliability levels and strong evidence of convergent validity as shown by Cronbach’s alpha 

and statistically significant λ x parameter estimates respectively. A more stringent confirmatory 

factor analysis using structural equation modeling reveals problems with discriminant validity, as 

evidenced by large λ x and Θ δ modification indices. Some scale items want to be associated with 

constructs that they are not supposed to be associated with. In addition, some error terms want 

to correlate with other error terms. This means that insufficient evidence exists to suggest that 

the dimensions of retail convenience are in fact different dimensions. Further studies must be 

undertaken to ensure that the dimensions of access, search, assisted search, transaction, and 

possession convenience are all distinct dimensions of convenience. 

The support of Hypothesis 2 suggests that online shoppers do benefit from access 

convenience more than traditional in-store shoppers do. One implication of this finding is that 

when given a choice between the two retail formats (online shopping or traditional in-store 

shopping) customers who value access convenience above all other types of convenience will 

often choose to shop online. Hypothesis 3 (that online shoppers benefit from search convenience 

more than traditional in-store shoppers) is also supported. When given a choice between the two 

retail formats, customers who value search convenience above all other types of convenience 

will also choose to shop online. 

268

From a marketing standpoint, the results of Hypothesis 2 and 3 suggest that retailers should 

try to better understand their customers; specifically, they should try to find out what type of 

convenience their customers want. It is possible that customers may value multiple convenience 

dimensions. For example, a customer may value both search convenience and possession 

convenience. One way to better satisfy this customer is to offer multiple channel outlets 

(maintain both a traditional brick-and-mortar store and an online store). By having multiple 

channel outlets, the retailer facilitates both search convenience and possession convenience. In 

other words, the customer can use the online store for his information search, then go to the 

traditional store and leave with his desired purchase. Offering multiple channels to reach 

customers offers ultimate convenience. 

Upon further review of hypothesis 4, the seemingly counter-intuitive results can be easily 

explained. It is true that online shoppers must wait longer for their purchases than in-store 

shoppers. But a post hoc examination of the item wordings that measured possession 

convenience for online shoppers reveals that this wait was not adequately captured. Instead, we 

asked if the order was delivered in a timely fashion. This captures satisfaction with possession, 

rather than how conveniently the possession actually took place. In subsequent scale refinement 

attempts, possession convenience might be captured more precisely if an objective measure was 

used. Suppose the question was worded, “How long did it take for you to receive the product(s) 

you ordered?” An in-store shopper would answer “zero” (the most convenient possession), and 

the longer one had to wait for his online order, the less convenient his possession would be. 

Conclusion 

This study is an initial attempt to delineate the different retail convenience dimensions 

for in-store and online shoppers. Several limitations must be mentioned. A convenience sample 

was used, and the respondents were all concentrated in one geographical area. Additionally, we 

269

obtained a small sample size of online shoppers. Scale refinement will greatly benefit from a 

national sample with equally large numbers of in-store and online shoppers. 

As mentioned above, future research should aim to establish discriminant validity of the 

convenience scales. Additional comparisons of in-store and online shoppers are also warranted. 

Future studies should investigate the time and effort expenditures for these two groups as well as 

examine possible antecedents to convenience-seeking behaviors. Constructs such as time 

pressure and shopping orientation may lead consumers to seek different types of convenience. 

Retailers would greatly benefit from such an understanding of their customers’ needs and 

motivations. 

270

References 

Anderson Jr. WT. Convenience orientation and consumption behavior. Journal of Retailing 

1972;48(3):49-72. 

Berry LL. The time-buying consumer. Journal of Retailing 1979;55(4):58-69. 

Berry LL, Seiders K, Grewal D. Understanding service convenience. Journal of Marketing 

2002;66(3):1-17. 

Berry LL. The old pillars of new retailing. Harvard Business Review 2001;79(4):131-137. 

Bettman JR, Johnson EJ, Payne JW. A componential analysis of cognitive effort in choice. 

Organizational Behavior and Human Decision Processes 1990;45(February):111-139. 

Bhatnagar A, Misra S, Rao HR. On risk, convenience, and Internet shopping behavior. 

Communications of the ACM 2000;43(11):98-110. 

Brown LG. The strategic and tactical implications of convenience in consumer product 

marketing. Journal of Consumer Marketing 1989;6(3):13-19. 

Brown LG. Convenience in services marketing. Journal of Services Marketing 1990;4(1):53- 

59. 

Bucklin LP. Retail strategy and the classification of consumer goods. Journal of Marketing 

1963;27(1):50-55. 

Burke RR. Technology and the customer interface: what consumers want in the physical and 

virtual store. Journal of the Academy of Marketing Science 2002;30(4):411-432. 

Childers T L, Carr CL, Peck J, Carson S. Hedonic and utilitarian motivations for online retail 

shopping behavior. Journal of Retailing 2001;77(4):511-535. 

Churchill GA. A paradigm for developing better measures of marketing constructs. Journal of 

Marketing Research 1979;16(1):64-73. 

Copeland MT. Relation of consumers’ buying habits to marketing methods. Harvard Business 

Review 1923;1(3):282-289. 

Crist B. Myths about convenience food costs. Journal of Marketing 1960;24(4):49-55. 

Darian JC, Cohen J. Segmenting by consumer time shortage. Journal of Consumer Marketing 

1995;12(1):32-44. 

DeVellis RF. Scale development: theory and applications (Applied Social Research Methods 

Series, Vol. 26). Newbury Park: Sage Publications, 1991. 

271

Dunn DT, Thomas CA, Lubawski JL. Pitfalls of consultative selling. Business Horizons 

1981;24(5):59-65. 

Enis BM, Roering KJ. Product classification taxonomies: synthesis and consumer implications. 

In: Theoretical developments in marketing. Chicago: American Marketing Association, 

1980. p. 186-189. 

Etgar, M. The household as a production unit. In: Sheth JN, editor. Research in marketing, vol. 

1. Greenwich, CT: JAI Press, 1978. p. 79-98. 

Foote NN. The Image of the Consumer in the Year 2000. Proceedings, Thirty-fifth Annual 

Boston Conference on Distribution 1963, 13-18. 

Forster J. Hungry for convenience. Business Week (1/14/2002); Issue 3765: 120. 

Gardner EH. Consumer goods classification. Journal of Marketing 1945;9(3):275-276. 

Green H. Happy holidays – for e-tailers, at least. Business Week (11/11/02); Issue 3807:44-45. 

Gross BL. Time scarcity: interdisciplinary perspectives and implications for consumer behavior. 

In: Sheth JN, Hirschman EC, editors. Research in consumer behavior. Greenwich, CT: 

JAI Press, 1987. p. 1-54. 

Hof RD. Don’t cut back now. Business Week (10/1/2001); Issue 3751:EB34. 

Hofacker CF. Internet Marketing, 3 rd ed. New York: John Wiley & Sons, Inc., 2001. 

Hoffman NP. May I help you?: customer intimacy versus customer empowerment in a retail context. 

Presented at the Third Annual Retail Strategy and Consumer Decision Research Seminar 2000, 

Orlando, FL, November. 

Holbrook MB, Howard JA. Frequently purchased nondurable goods and services. In: Ferber R, 

editor. Selected aspects of consumer behavior: a summary from the perspective of 

different disciplines. Washington D.C.: National Science Foundation, Directorate for 

Research Applications, Research Applied to National Needs, 1977. p. 189-222. 

Jacoby J, Szybillo GJ, Berning CK. Time and consumer behavior: an interdisciplinary overview. 

Journal of Consumer Research 1976;2(4):320-339. 

Kaish S. Cognitive dissonance and the classification of consumer goods. Journal of Marketing 

1967;31(4):28-31. 

Kaufman-Scarborough C. , Lindquist JD. E-shopping in a multiple channel environment. 

Journal of Consumer Marketing 2002; 19(4):333-350. 

272

Kelley EJ. The importance of convenience in consumer purchasing. Journal of Marketing 

1958;23(1):32-38. 

Kotler P, Zaltman G. Social marketing: an approach to planned social change. Journal of 

Marketing 1971;35(3):3-12. 

Lovelock CH. Classifying services to gain strategic marketing insights. Journal of Marketing 

1983;47(3):9-20. 

Morganosky MA, Cude BF. Consumer response to online grocery shopping. International 

Journal of Retail and Distribution Management 2000;28(1):17-26. 

Murphy PE, Enis BM. Classifying products strategically. Journal of Marketing 1986;50(3):24- 

42. 

Prest AR, Turvey R. Cost-benefit analysis: a survey. Economic Journal 1965;75(300):683-735. 

Reilly MD. Working wives and convenience consumption. Journal of Consumer Research 

1982;8(4):407-418. 

Seiders K, Berry LL, Gresham LG. Attention, retailers! How convenient is your convenience 

strategy? Sloan Management Review 2000;41(3):79-89. 

Seiders K, Voss GB, Grewal D, Godfrey AL. Customer evaluation of service convenience: an 

empirical investigation. Proceedings, American Marketing Association Winter 

Educators’ Conference, (Winter 2003), 163-164. 

Shop.org. The multi-channel retail report 2001. Washington D.C.: National Retail Federation, 

2001. 

Spector, PE. Summated Rating Scale Construction: An Introduction. Newbury Park: Sage 

Publications, 1992. 

Szymanski DM, Hise RT. E-satisfaction: an initial examination. Journal of Retailing 

2000;76(3):309-322. 

Wolfinbarger M, Gilly MC. Shopping online for freedom, control, and fun. California 

Management Review 2001;43(2):34-55. 

Youngdahl WE, Kellog DL. The relationship between service customers’ quality assurance 

behaviors, satisfaction, and effort: a cost of quality perspective. Journal of Operations 

Management 1997;15(1):19-32. 

273

Table 1 

Dimensions of Convenience 

General Convenience Definition 

(Brown 1989, 1990) 

1. Time Products/services may be provided at a time that is most 


2. Place Products/services may be provided in a place that is more 


3. Acquisition Firms may make it easier for the customer, financially and 

otherwise, to purchase their products/services 

4. Use Products/services may be made more convenient for the 

customer to use 

5. Execution Perhaps the most obvious convenience is simply having some 

one else provide the product/service for the consumer 

Service Convenience (Berry, 

Seiders, and Grewal 2002) 

1. Access convenience Involves consumers’ perceived time and effort expenditures to 

initiate service delivery 

2. Decision convenience Involves consumers’ perceived time and effort expenditures to 

make service purchase or use decisions 

3. Transaction convenience Involves consumers’ perceived expenditures of time and effort 

to effect a transaction 

4. Benefit convenience Involves consumers’ perceived time and effort expenditures to 

experience the service’s core benefits 

5. Post-benefit convenience Involves consumers’ perceived time and effort expenditures 

when reinitiating contact with a firm after the benefit stage of 

the service 

Retail Convenience (Seiders, 

Berry, and Gresham 2000) 

1. Access convenience The speed and ease with which consumers can reach a retailer 

2. Search convenience The speed and ease with which consumers can identify and 

select products they wish to buy 

3. Transaction convenience The speed and ease with which consumers can effect or amend 

transactions 

4. Possession convenience The speed and ease with which consumers can obtain desired 

products 

274

Dimension Items for In-Store Shoppers 1 

Table 2 

Scale Items Measuring Retail Convenience 

Items for Online Shoppers 1 

Access 1. The store was easy to get to. 1. The website was easy to find. 

2. The store had convenient hours. 2. I could order any time I wanted. 

3. Parking was reasonably available. 3. I could order from remote locations 

4. I was able to get to the store’s location 

quickly. 

(e.g., home, work, etc.) 

4. I was able to find the website quickly. 

Search 1. The store was well-organized. 1. The website was well-organized. 

2. I could easily find what I was looking 2. It was easy to find what I was looking 

for. 

for. 

3. The store was neat. 3. It was easy to navigate the website. 

4. The store was clean. 4. The website provided useful 

information. 

5. It was easy to get the information I 

needed to make my purchase decision. 

Help with Search 1. A salesperson was readily available to 

help me. 

2. The salesperson offered good advice. 

3. The salesperson helped me find exactly 

what I was looking for. 

4. The salesperson was knowledgeable. 

5. The salesperson gave me useful 

information. 

6. I thought the salesperson was welltrained. 

Transaction 1. The store has a fast checkout. 1. The checkout process was fast. 

2. My purchase was completed easily. 2. My purchase was completed easily. 

3. I was able to complete my purchase 3. It didn’t take a long time to complete 

quickly. 

4. I didn’t have to wait to pay. 

5. It didn’t take a long time to complete 



Possession 1. I got exactly what I wanted. 1. I got exactly what I wanted. 

2. It took a minimal amount of effort on 2. It took a minimal amount of effort on 



3. I got what I wanted when I wanted it. 3. My order was delivered in a timely 

fashion. 

4. I was properly notified of my order 

status. 

1All items were measured on a 7-point scale anchored by Strongly Agree and Strongly Disagree. 

275

Table 3 

Retail Convenience Scale Results for In-Store Shoppers 

Dimensio 

n 


1 2 3 4 5 λx estimate t-value SMC 

Access1 0.7759 0.87 

7 

0.89 16.56 0.80 

Access2 .077 

1 

0.73 12.48 0.53 

Access3 .063 

1 

0.57 9.26 0.33 

Access4 0.85 

7 

0.79 13.95 0.62 

Search1 0.8860 0.72 

6 

Search2 0.62 

2 

Search3 0.85 

8 

Search4 0.82 

6 

Help1 0.9503 0.77 

3 

Help2 0.88 

1 

Help3 0.79 

9 

Help4 0.92 

3 

Help5 0.90 

5 

Help6 0.87 

1 

Transact1 0.9609 0.88 

5 


7 


1 


0 


3 

0.84 15.27 0.71 

0.76 13.28 0.58 

0.78 13.62 0.61 

0.73 12.43 0.53 

0.75 13.74 0.56 

0.90 18.15 0.81 

0.84 16.26 0.71 

0.94 19.71 0.89 

0.94 19.45 0.88 

0.88 17.33 0.77 

0.87 17.24 0.76 

0.95 20.05 0.90 

0.98 21.07 0.95 

0.84 16.27 0.70 

0.95 20.19 .091 

Possess1 0.7649 0.761 0.76 12.39 0.57 

Possess2 0.663 0.67 10.74 0.45 

Possess3 0.778 0.72 11.63 0.52 

n=241 

276

Dimensio 

n 

Table 4 

Retail Convenience Scale Results for Online Shoppers 


1 2 3 4 λx estimate t-value SMC 

Access1 0.9298 0.866 0.86 8.26 0.73 

Access2 0.735 0.81 7.65 0.66 

Access3 0.852 0.93 9.59 0.87 

Access4 0.903 0.92 9.34 0.85 

Search1 0.9033 0.537 0.86 8.31 0.75 

Search2 0.508 0.87 8.46 0.76 

Search3 0.738 0.88 8.65 0.78 

Search4 0.866 0.66 5.61 0.43 

Search5 0.850 0.65 5.55 0.42 

Transact1 0.9722 0.921 0.96 10.22 0.93 

Transact2 0.914 0.97 10.35 0.94 

Transact3 0.918 0.95 10.04 0.91 

Possess1 0.9110 0.800 0.85 8.12 0.73 

Possess2 0.669 0.83 7.80 0.69 

Possess3 0.845 0.87 8.45 0.76 

Possess4 0.897 0.84 7.98 0.71 

n=67 

277

Table 5 

Correlations, Means, and Standard Deviations of Scale Items for In-store Shoppers 

Item A1 A2 A3 A4 S1 S2 S3 S4 H1 H2 H3 H4 H5 H6 T1 T2 T3 T4 T5 P1 P2 P3 

Acc1 1 

Acc2 .63 1 

Acc3 .49 .50 1 

Acc4 .73 .54 .40 1 

Sear1 .37 .41 .40 .31 1 

Sear2 .26 .29 .25 .21 .63 1 

Sear3 .23 .32 .30 .16 .67 .60 1 

Sear4 .27 .35 .34 .21 .63 .53 .80 1 

Help1 .20 .12 .10 .18 .38 .38 .38 .33 1 

Help2 .17 .14 .06 .10 .39 .41 .39 .35 .70 1 

Help3 .18 .17 .12 .11 .41 .48 .42 .37 .68 .81 1 

Help4 .19 .17 .07 .10 .40 .38 .38 .38 .70 .83 .78 1 

Help5 .18 .15 .07 .08 .35 .36 .39 .37 .66 .85 .75 .90 1 

Help6 .15 .11 .06 .14 .36 .40 .41 .38 .69 .76 .72 .84 .82 1 

Tran1 .21 .28 .23 .12 .37 .33 .22 .31 .25 .31 .35 .27 .26 .22 1 

Tran2 .21 .24 .22 .13 .33 .43 .33 .34 .29 .34 .42 .31 .29 .32 .83 1 

Tran3 .18 .22 .23 .11 .32 .42 .32 .32 .26 .31 .39 .27 .27 .27 .84 .93 1 

Tran4 .12 .12 .22 .03 .31 .38 .26 .36 .27 .28 .35 .28 .23 .25 .79 .77 .81 1 

Tran5 .18 .20 .17 .10 .29 .41 .30 .29 .26 .32 .39 .28 .25 .25 .83 .90 .93 .82 1 

Pos1 .07 .10 .03 .05 .27 .39 .27 .22 .39 .46 .56 .42 .44 .43 .30 .41 .36 .30 .37 1 

Pos2 .27 .24 .20 .16 .35 .36 .31 .30 .23 .30 .34 .20 .25 .21 .41 .46 .49 .40 .46 .49 1 

Pos3 .21 .25 .16 .13 .35 .42 .32 .33 .33 .29 .39 .24 .30 .30 .32 .39 .37 .29 .32 .57 .47 1 

Mean 5.6 5.9 5.8 5.3 5.8 5.8 5.9 6.0 6.0 5.7 5.8 6.0 5.8 5.9 5.1 5.3 5.1 5.2 5.1 6.0 5.0 5.8 

SD 1.4 1.2 1.3 1.7 1.1 1.2 1.2 1.0 1.4 1.5 1.5 1.3 1.3 1.3 1.5 1.6 1.6 1.7 1.7 1.3 1.8 1.5 

n=241 

Table 6 

Correlations, Means, and Standard Deviations of Scale Items for Online Shoppers 

Item A1 A2 A3 A4 S1 S2 S3 S4 S5 T1 T2 T3 P1 P2 P3 P4 

Acc1 1 

Acc2 .64 1 

Acc3 .77 .79 1 

Acc4 .85 .72 .86 1 

Sear1 .56 .66 .62 .59 1 

Sear2 .51 .67 .67 .55 .76 1 

Sear3 .47 .66 .60 .52 .78 .74 1 

Sear4 .29 .51 .38 .37 .52 .51 .71 1 

Sear5 .29 .46 .33 .28 .48 .59 .67 .77 1 

Tran 

1 

.24 .53 .34 .33 .43 .58 .48 .38 .42 1 

Tran 

2 

.16 .48 .26 .21 .39 .55 .49 .46 .54 .93 1 

Tran .15 .42 .27 .22 .37 .50 .48 .43 .50 .92 .93 1 

3 

Pos1 .44 .45 .46 .42 .62 .65 .50 .30 .42 .54 .47 .43 1 

Pos2 .41 .63 .56 .48 .60 .66 .51 .41 .43 .61 .53 .55 .74 1 

Pos3 .47 .47 .59 .52 .51 .59 .49 .20 .19 .38 .29 .29 .71 .70 1 

Pos4 .41 .35 .51 .43 .53 .48 .41 .19 .16 .31 .24 .24 .72 .65 .81 1 

Mean 6.5 6.6 6.5 6.6 6.2 6.3 6.2 6.3 6.3 5.7 5.7 5.6 6.4 5.9 6.2 6.4 

278

SD .89 .70 .80 .82 .91 .98 .91 .85 .88 1.2 1.4 1.3 .80 1.2 1.2 1.1 

n=67 

279

Error! 

Figure 1 

Retail Convenience for In-Store and Online Shoppers 

In-Store Retail 

Convenience 

Access Search Assisted 

Search 

Transaction Possession 

Online Retail 

Convenience 

Access Search Transaction Possession 

280

What influences customer reaction to incompatibility? An empirical investigation 

Abstract 

This paper seeks to understand the factors that influence customer reaction 

to incompatible behavior of fellow customers in context of foodservice business. 

Based on extensive literature review, we develop hypotheses about the direction 

of relationship between controllable and uncontrollable psychological variables 

and individual reactions to incompatibility. Using experimental design technique 

we confirm these hypotheses. Mood, expectations, and perceived control over 

outcome were found to be three most important factors influencing the intensity 

of reaction towards incompatibility. Using Taguchi experimental design technique, 

we further develop optimal combination of alternate levels of controllable and 

uncontrollable design factors that minimize the negative impact of uncontrollable 

factors. Managerial implications of research are discussed. 

Introduction 

281

Customer-to-customer interaction is known to be an important aspect of 

overall service encounter as the behavior of fellow customers can impact overall 

satisfaction and possible future purchase intentions (Parker and Ward 2000, 

Bitner et al. 1997, Swinyard 1993). Customer satisfaction is a recursive process 

where satisfaction/dissatisfaction with a previous service encounter influences 

future patronage intentions. Four this process to work, consumers rely on some 

kind of recall from previous service experience. This recall tends to be biased in 

the direction of momentary emotions associated with the event being recalled 

(Machleit et al. 2000). Therefore the momentary emotions felt during service 

encounters such as negative response to incompatible behavior of fellow 

customers are likely to influence recall and future intention to purchase. 

In situations where customers have to share space and time, customer-to- 

customer contact -both verbal and non-verbal- is inevitable. It is also inevitable 

that customers will find themselves compatible with some customers and 

incompatible with others (Martin 1996, Martin and Partner 1989). In cases where 

incompatibility is experienced, the response is expected to vary among 

customers. Martin (1989 pp. 12) points out “some customers may view certain 

behaviors in particular service environments as intolerable, yet other customers 

may not be disturbed”. The reaction towards incompatible behavior of fellow 

customers can, therefore, be characterized as individual specific. Since the 

reaction to incompatibility can range from indifference to negative word of mouth 

(Martin 1989), It is important for marketers to understand the factors that 

moderate the intensity of this reaction. 

282

Management of in compatibility has traditionally been treated as an 

uncontrollable phenomenon that needs to be avoided rather than something that 

should be managed operationally. Some of the factors that are expected to 

moderate the intensity of reaction to incompatibility such as mood (Swinyard 

1993), involvement (Swinyard 1993), and personal values (Schwartz 1992) are, 

unfortunately, beyond the control of marketing managers. That is, perhaps, the 

reason for emphasis on avoiding incompatibility through effective management of 

different customer segments. Articulation of a clear positioning statement which 

enables customers to self select businesses where they will feel comfortable in 

the company of their fellow customers is expected to help avoid incompatible 

situations. 

Complete avoidance of incompatibility will, however, require extreme 

segmentation which may not be a practical option for all businesses. It has been 

suggested that for most of the businesses to grow, they need to serve diverse 

groups of customers which show different levels of tolerance to incompatible 

behaviors (Martin 1996). Incompatibility, therefore, seems unavoidable. In 

addition to effective segmentation incompatibility also needs to be dealt at an 

operational level. Managers may use both on site solutions such as physical 

separation, enforcement of policies designed to control unacceptable behavior, 

and advance communication with customers about the possibility of encountering 

some undesirable behaviors. 

For an active management of incompatibility we need to accomplish two 

tasks. First, we need to identify the factors that influence the intensity of reaction 

towards undesirable public behaviors and understand their relative contributions. 

283

Second, we need to find out the optimal combination of controllable factors that 

would minimize the negative impact of uncontrollable factors. In marketing 

literature, there is an apparent lack of information on either of these issues. This 

article is an attempt to bridge this gap of information. 


Research on the role of customers in service encounters has traditionally 

been limited to the study of interaction between customers and service 

employees. The influence of customer-to-customer interaction on service 

encounter quality, satisfaction, and future purchase intentions has largely been 

ignored in service marketing literature. This omission becomes clear when we 

look at prominent service quality measurement scales such as SERVEQUAL and 

SERVPREF which fail to capture the effect of customer-to-customer interaction. 

Although some of the researchers (Baker 1987) do discuss a social dimension of 

tangible quality that represents organization’s concern for the people in the 

environment such as customer, employees and non-customers, measurement 

researchers in general have failed to incorporate this concern in to existing 

service quality scales. 

Research in the area of customer-to-customer interactions can be 

grouped in into four categories i.e. identification and grouping of public behaviors, 

role adoption, impact of demographic and individual factors on reaction to 

incompatibility, and compatibility management. Grove ad Fisk (1997) and Martin 

(1996) developed categories of positive and negative public behaviors at 

business places. Grove and Fisk employed critical incidence technique to drive 

the dimensions while Martin used exploratory factor analysis to extract the same. 

284

A lot of commonalities can be found between seven factor solution by Martin and 

six categories presented by Grove and Fisk. In fact most of the Grove and Fisk’s 

six categories can easily be fitted into Martin’s seven factor solution. Since the 

tolerance for public behaviors is situation and individual specific (Martin 1996), 

the classification of public behavior can serve as a good basis for customer 

segmentation. This classification also alerts service managers to identify and 

control the negative public behaviors and encourage the positive behaviors. 

Research on roles fellow customers play is also very limited. Parker and 

Ward (2000), Bitner et. al. (1997), and McGrath and Otnes (1995), describe roles 

that customers may play in improving service experience of other customers. 

These studies identify two basic roles customers adopt when interacting with 

each other; help seeker and help provider. By identifying typical scripts 

associated with these roles, these studies stress the huge opportunities that 

service managers need to exploit by planning and encouraging interactions 

among customers. 

Research on the factors influencing customer perception of incompatibility 

is limited to two studies only. Martin and partner (1989) in an exploratory study 

propose that incompatibility can arise because of demographic, social and 

cultural differences. These differences may include different preferences, 

different purchase goals or sought benefits, stereotyping, beliefs, attitude and 

cultural backgrounds, differences in past experience with service and other 

customers’ physical characteristics and medical condition. These propositions 

were based on interviews and customer surveys. Grove and Fisk (1997) while 

empirically testing the impact of individual demographic differences on the 

285

esponse to the behavior of other customers found that most of the demographic 

differences such as age, education, income and gender are not related to our 

behavior towards fellow customers. They however found that marital status is 

positively related to dissatisfaction with the behavior of fellow customers. 

Since most of the demographic variables fail to explain much of the 

variation in customers’ reactions towards incompatibility, we need additional 

variables to explain the variation. Another set of variables, personal psychological 

variables, seems promising in improving our understanding of incompatibility. 

Variables such as mood states, perceived control, prior expectations, 

involvement and personal values has been demonstrated to be strongly linked to 

consumer purchase intentions (Machleit et al. 2000, Swinyard 1997, Hui and 

Bateson 1991). We believe that these variables, in addition to having a direct 

effect on purchase intentions, also directly impact the intensity of a consumers’ 

reaction to the incompatibility of their fellow customers which in turn contributes 

to future purchase intentions and re-patronage. Although Martin (1989) suggests 

the possible impact of personal psychological variables such as individual beliefs, 

attitudes, personal values on our reactions to customer incompatibility, his 

propositions have not been tested. 

In providing guidelines for incompatibility management, researchers have 

predominantly opted for segmentation as the preferred method to avoid 

unpleasant situations (Martin 1996, Martin and Partner 1991, Martin and Partner 

1989). Customer education on their roles in achieving satisfaction, clear 

communication about the acceptable behaviors (Grove and Fisk 1997, Bitner et 

al. 1997), realistic expectation setting (Bitner et al. 1997), and efficient 

286

management of space (Martin and Partner 1997) has also been suggested to 

reduce the negative impacts of incompatibility. No empirical study, however, is 

available that tests the impact of expectation management, enforcement of the 

behavioral code of conduct and physical space management (which is expected 

to improve customers’ perception of control) on reaction to incompatible 

situations. 

Literature review establishes two basic gaps. First, the impact of personal 

psychological variables on reaction towards incompatibility has not been 

considered. Second, no empirical testing has been done on neither the impact of 

uncontrollable factors such as mood, values or involvement and/or the impact of 

controllable factors such as expectation communication and physical 

management of the space on reaction to incompatibility; neither the interactions 

between these factors. 

In this study, we attempt to fill both of these gaps. We start by developing 

hypotheses about the impact of both controllable and uncontrollable factors on 

customers’ reaction to incompatible situations. We proceed further by testing 

these hypotheses using an experimental design setting. 

Hypotheses 

Mood can be conceptualized as mild, transient feeling states that are 

subjectively perceived by individuals. Mood states may affect cognitive 

processes such as evaluation, memory and decision making (Gardner 1985). It 

has often been conceptualized as constituting of two independent bipolar positive 

and negative feeling states (Isen 1984). A positive mood seems to make one 

kinder, generous, more resistant to temptations and more willing to delay self 

287

ewards (Swinyard 1993). A customer in positive mood is expected to 

demonstrate a more positive evaluation of service encounter than those with 

negative or positive mood (Knowles et al. 1993). In other words, customers in 

good mood will display greater tolerance towards negative impacts of 

incompatibility. Negative moods on the other hand produce greater dislike for 

fellow customers, greater vulnerability to provocation and lesser willingness to 

forgive (Gardner 1985). Alternatively stated, a customer in negative mood can 

strongly react to even small incompatible behaviors. 

H1 : Customers in good mood will show higher tolerance for 

incompatibility and customers in bad mood will show low tolerance 

for incompatibility. 

Recent research suggests there are two levels of service expectations i.e. 

desired and adequate (Kalra 1991 and Boulding et al. 1993). Desired 

expectations represent the level of service a consumer hopes to receive, while 

adequate expectations, a lower level of expectations, relate to what consumers 

deem an acceptable level of performance. Desired expectations are known to 

remain relatively stable over time, while adequate performance expectations may 

vary more widely. Between these two service quality expectation levels is a zone 

of tolerance. The zone of tolerance describes the latitude in customer’s 

willingness to accept deviation from desired standards. This zone of tolerance will 

be different for different services (Gwynne and Devlin 2000). For example 

patients in waiting room to see a doctor will expect and accept a low level of 

compatibility where as a customer in his/her favorite bar may not expect and 

accept a high level on incompatibility. It follows then that if the level of customer 

incompatibility lies within the zone of tolerance, customer evaluation of service 

288

may not be adverse. Expected incompatibility may lead to a reduction in the 

intensity of reaction while the unexpected incompatibility may increase the 

intensity of reaction. 

H2 : Prior expectations of incompatibility will increase tolerance for 

incompatibility. 

Control refers to the ability to regulate one’s own affairs. It has often been 

defined as the need to demonstrate one’s competence, superiority and mastery 

over the environment (White 1959). Literature makes a distinction between 

behavioral, cognitive and decisional control (Averill 1973). Behavioral control – 

also referred to as control over activities—“refers to availability of a response 

which may directly influence or modify the objective characteristic of an event” 

(pp.287). Decisional control or control over outcome refers to “the opportunity to 

choose among various courses of action” (pp.287). The cognitive control refers to 

“the way in which the event is interpreted, appraised, or incorporated into a 

cognitive plan” (pp.287). In a restaurant situation control over activities may be 

though as the ability to regulate improper public behaviors of fellow customers by 

asking manager to enforce rules of conduct while control over outcome can be 

operationalized as the ability to escape an undesirable situation by moving to 

another section of the restaurant. Perception of high control has been shown to 

positively impact the tolerance of pain and frustration, and self-report of distress 

and anxiety (Hui and Bateson 1991). 

H3 : Higher perceived control over service activities will increase 

tolerance for incompatibility. 

H4 : Higher perceived control over service outcome will increase 

tolerance for incompatibility. 

289

Values may be defined as beliefs that pertain to desirable end states or 

modes of conduct, transcend specific situations such as behavior of fellow 

customers, guide selection or evaluation of behavior, and are ordered by 

importance in relation to one another to form a system of value priorities. 

Schwartz (1992) provides a typology of values, where the values are arranged 

according to the type of motivational goal they express. These ten value types 

represent two basic bipolar dimensions; openness to change versus 

conservation and self transcendence versus self-enhancement. In context of 

reaction to customer incompatibility, the bipolar dimension of conservation 

versus openness to change seems very relevant. Conservative individuals are 

very concerned with security, conformity, and tradition. These individuals are 

expected to display low tolerance for incompatibility. On the other hand 

individuals displaying openness are focused on hedonism, stimulation and self- 

direction and are expected to show greater tolerance for customer 


H5 : Customers with resultant conservatism will show low tolerance for 


H6 : Customers with resultant openness will show high tolerance for 


The intensity of customer involvement in purchase decision is also 

expected to moderate reaction to customer incompatibility. The information 

processing theory in consumer behavior literature suggests that the consumer is 

an intelligent, rational, problem-solver who actively seeks and uses information to 

evaluate the various alternatives or choices (Zaichowsky 1985). The extent of 

290

consumer involvement varies with the significance or degree of importance of the 

decision to the consumer (Engel and Blackwell 1982). Depending on the degree 

of involvement in a purchase decision, consumer decision making has been 

classified as being high or low involvement. In a low-involvement purchase 

decision, because the purchase does not have much significance to the 

consumer, he or she does not extensively search for information, and rarely 

evaluates alternatives or choices before making the purchase decision. In order 

to reduce the risk associated with high-involvement decisions, consumer 

searches for more information and spend more time searching for the right 

selection. Because of this high risk component, high involvement service 

purchasers are expected to show low tolerance for customer incompatibility that 

could interfere with desired service outcome. On the other hand, low involvement 

service purchasers are expected to show greater tolerance for customer 


H7 : Customers with high involvement will show low tolerance for 


Sources of Data 

Two hundred and ten respondents were first self-administered Personal 

value system questionnaire. Respondents were then separated into groups of 

open to change and conservative respondents. 

Methodology 

Two different experiments were conducted. The purpose of the first 

experiment was to confirm whether the factors being considered in the 

experiment and some of the first order interactions had significant effect on 

reaction to incompatible situations. The second experiment was conducted to 

291

find out the optimal combination of controllable and uncontrollable factors would 

minimize the negative impact of uncontrollable factors on reaction to 

incompatible situations. 

For understanding main and interaction effects, a 

6 2 

2 − fractional factorial 

iv 

design with sixteen treatments, as shown in figure 1, was used to create service 

scenarios or profiles for understanding the impact of antecedents on reaction 

towards customer incompatibility. Since it was a resolution four design, no main 

effect was confounded with first order interactions. This means that not only main 

effect estimates were clean but we could also calculate some of the first order 

interactions rather cleanly. 

--------------------------------------------------------------------------------------------------- 

Place Figure 1 about here 

--------------------------------------------------------------------------------------------------- 

Three hundred and twenty respondents self administered 27 item 

Schwartz value system questionnaire for measuring conservative versus 

openness to change orientation. Based on the results of the survey, respondents 

were segregated into two groups of conservative and open to change. Following 

Weyant (1978), mood of each group was manipulated through false feedback. 

Half of the respondents in each group were given two different types of anagrams 

to be solved. Anagrams are the words, from whose alphabet another word can 

be created. One set of respondents in each group was handed anagrams which 

were supposed to easier to solve while the other group got a difficult set. Each 

anagram test contained 25 questions. Subjects were given 15 minutes to 

complete as many anagrams as possible. After the end of 15 minutes 

respondents read the false statistics about the tests. Respondents in the groups 

that received easy anagrams were told that they had done exceedingly bad and 

292

their results were well below the average for these kinds of tests, thereby placing 

them in bad mood. Group that received difficult anagrams were told that they had 

done exceedingly well and their results were far above the average for these 

kinds of tests, thereby placing them in good mood. 

For treatment of control over activities, control over outcome, 

expectations, and involvement, each respondent received one of the sixteen 

combinations (in form of a video movie) and recorded his/her reaction to 

customer incompatibility on a scale of 1-6. Instead of using written scenarios, 

video films were developed to operationalize the design factors of involvement, 

perceived control and expectations. We believe that video films provided 

respondents were in restaurant reality and their responses were expected to 

more realistic than would have been with written scenarios. At the end, 

manipulation measurement checks were made for mood, perceived control and 

involvement. 

For experiment to find out the optimal combination of controllable and 

uncontrollable factors, half fractions of 2 3-1 were used as outer and inner arrays in 

a Taguchi style experiment (see figure 2). The inner array is composed of 

controllable factors and the outer array consists of uncontrollable factors. In this 

arrangement each combination of controllable factors was contrasted with all 

combination of uncontrollable factors providing us with detailed information on all 

possible impacts of uncontrollable factors. 

--------------------------------------------------------------------------------------------------- 

Place Figure 2 about here 

--------------------------------------------------------------------------------------------------- 

293

Taguchi recommends the use of signal to noise ratio (S/N ratio) as the 

basis for selecting the optimal combination, where the term signal represents the 

controllable component while the term noise represents the uncontrollable 

component. This ratio, infact, measures the ability to perform according to 

specifications in varying conditions. The objective of S/N ratio is to determine an 

optimum set of operating conditions that simultaneously maximizes the objective 

quality function while reducing the variation in output quality. 

Taguchi recommends three types of S/N ratios: Larger the better 

(maximizing the desired response variable), smaller the better (minimizing the 

undesirable response variable) or minimal is best (attaining a predefined level). 

In this case, because we were trying to maximize choice probabilities, the 

appropriate performance statistic deemed to be the “smaller the better”. It is 

given by the following formula: 

SN 

s 

⎡1 

n 

= − 10 log ⎢ ∑ Y 

⎣n 

i = 1 

2 

i 

⎤ 

⎥ 

⎦ 

where yi is the response variable which in our case is the reaction to incompatible 

situation, and n is the number of observations. 

Measures 

Mood: Mood scale was adapted from Peterson and Sauber (1983). Using a 

seven point likert type scale, mood was measured by four summed bipolar items. 

Four bipolar items were sad/happy, good mood/bad mood, irritable/pleased and 

depressed/cheerful. 

Involvement: We used a modified form of involvement scale used by Gore and 

Madhavan (1994) for measuring consumer involvement in nonprescription 

294

medicine purchase decisions. A five item seven point likert scale was used. 

These items represented the various theoretical propositions about the 

involvement construct such as search for information, evaluation of product 

alternatives, and perception of differences among the various brands. 

Perceived Control: Following Hui and Bateson (1991), we used a semantic 

differential combination of Meharbian and Russel’s (1974) scales of dominance 

and Glass and Singer’s (1972) scales of helplessness as a proxy for the 

construct of perceived control. 

Expectations: Prior expectation of customer compatibility were measured by 

asking respondents to rate their expectation regarding compatibility on seven 

point (0-6) scale where 0 represents no compatibility incompatibility and 6 

represents high compatibility. 

Personal Values: For measuring personal values, a twenty seven item scale 

(Schwartz, 1992) was used. These questions represented bipolar dimensions of 

conservatism versus openness to change Schwartz value system has been used 

and found reliable in many different parts of the world including surveys in forty- 

four countries (Schwartz and Sagiv 1995). 

Compatibility Threshold: We used six point customer compatibility 

dissatisfaction reaction scale developed by Martin (1989). The scale ranges from 

a tolerable behavior of “would not effect me one way or the other” to an extremely 

sever reaction “would bother me enough that I would never return”. 

Results 

Table 1 presents the estimated marginal means for the reaction to 

incompatible situation for both levels across six factors. Lower marginal means 

295

epresent higher tolerance of incompatibility while higher marginal means 

represent lower tolerance. All results were in expected direction. Data supports 

all seven hypotheses. Means for good and bad mood (2.63 and 3.70 

respectively) provide strong support for hypothesis 1. The range of marginal 

means (the difference between means for two levels) signifies the relative 

intensity of the impact of design factors on response variable. Larger is the 

range, higher is the impact. Range for mood is the third highest among the six 

factors. Data also provide strong support for hypothesis 2, where marginal means 

for expectation of incompatibility and no expectation were 2.58 and 3.78 and had 

the second highest range (1.20). Hypotheses 3 about the impact of perception of 

control on activity received only moderate support (3.38 and 2.97 for perception 

of control over activities and no control respectively with a small range of 0.41). 

Hypothesis 4 received the strongest support (2.55 and 3.80 for control and no 

control over outcome with a range of 1.25). Hypotheses 5 and 6 received 

moderate support (2.92 and 3.43 for open to change and conservative 

orientations with a small range of 0.48). Similarly hypothesis 7 was also 

supported moderately with means of 2.90 and 3.46 for low and high involvement 

with a range of 0.56. 

Place table 1 about here 

Factorial analysis of variance was conducted to investigate the impact of 

main and interaction effects on reaction to customer incompatibility. Model seems 

to fit the data well (Adjusted R 2 =0 .893). ANOVA results presented in table 2, 

showed a significant effects for control over outcome (F = 322, p

(F=297, p

estaurant where they can escape incompatibility; communicate to them in 

advance about the probability of encountering incompatible behavior then even a 

conservative customers in bad mood and high involvement in restaurant selection 

will not react adversely to incompatibility. 

Discussion 

Place table 3 about here 

Purpose of this study was twofold; to explore the antecedents of customers’ 

reaction to incompatibility and to develop operational guidelines for practitioners 

to manage incompatibility at foodservice outlets. Although the factors selection 

and testing was based on review of literature that assumed these to be linked to 

incompatibility, it was important to test these commonly held assumptions. By 

testing hypotheses about this linkage we were able to determine the direction of 

these linkages. Although the model fit statistics are fairly high (adjusted R 2 =0 

.893) we feel that the model tested can not be considered as complete. Some of 

the factors that may have an impact on incompatibility such as educational level 

of respondent and price of the meal were not included in the experiment because 

of the increasing response burden on the respondents. Even the six factor 

experiment can be considered as one with high response burden and we tried to 

lighten up this burden by using video films in place of traditional written 

scenarios. 

Although the testing of hypotheses and information obtained from these were 

important in understanding the phenomenon of incompatibility, it unfortunately 

did not provide relevant guidelines in terms of managing incompatibility at the 

business place. It is obvious that in most of the cases, incompatibility at business 

place is inevitable. Good segmentation can help reduce incompatibility but it can 

298

not eliminate it. Further, extreme segmentation has its own shortcomings. For 

most of the businesses to grow, they have to attract diverse group of customers 

and diversity inevitably results in some form of incompatibility. The emphasis, 

therefore, has to be on managing on site compatibility. 

Results of experimental design on the antecedents of incompatibility 

indicated that only one of the three most important factors, perception of control 

over outcome, was under the control of operating manages. Customers mood 

was a beyond the control of managers, while customers prior expectations could 

only be considered semi controllable. If two of the three most important factors 

contributing to reaction towards incompatibility were beyond the control of 

operating managers, then the next logical question is what can be done about 

the impact of uncontrollable factors? Stated alternatively, what the optimal 

combination controllable and uncontrollable factor levels that will minimize the 

severity of customer reaction towards incompatibility? 

The answer to this question was obtained by conducting a Taguchi 

experiment. Results of Taguchi experiment suggested that if we provide 

customers with an option to escape the incompatibility by moving them to 

another section and alert them to the possibility of encountering incompatible 

situation through effective communications then the adverse impact of bad mood, 

conservatism and high involvement can be minimized. Enforcing the rules for 

proper code of conduct – perceived control over activities — alone, without 

moving customers to another section – perceived control over outcome – will not 

minimize the severity of reaction to incompatibility. 

299

Three significant interaction terms also provide useful information. A 

significant interaction between control over activities and mood suggests 

customers in bad mood will have greater desire for enforcing the rules of conduct 

regarding public behaviors. Similarly significant interaction between expectations 

and involvement suggests the customers with high involvement in the selection 

of restaurant such as those going for a romantic dinner have low expectations of 

incompatibility and therefore are expected to react severely to an incompatible 

situation. Perhaps, the most significant interaction was found to be between 

involvement and control over outcome. A higher involvement in the restaurant 

selection enhances the desire for greater control over outcome when faced with 

incompatible situation. Significant interaction of involvement with desired control 

over outcome and prior expectation of incompatibility suggests the threat of 

severe reaction to incompatible will be higher for full service dinning 

establishment which can generally be characterized as destinations that require 

high customer involvement than fast food restaurants than can generally be 

characterized as low involvement destinations. 

Using video films instead of traditional written scenario was another 

novelty of this research. Written scenarios at times may not be able to fully 

describe the situation in which respondent has to make a decision. The use of 

video film was not only expected to fully describe the situation in reality but also 

to reduce the response burden where respondents can watch the film and place 

themselves in a character’s shoes. Imaging a situation without the help of 

pictures will involve greater response burden. 

References 

300

Averill, J. R. (1973), “Personal control over aversive stimuli and its relationship to stress” 

Psychological Bulletin, 80, 286 –303. 

Baker, Julie (1987), “The Role of the Environment in Marketing Services: The Consumer 

Perspective,” in The Services Challenge: Integrating for Competitive Eds. John A.Cecil et al. 

Chicago, IL: American Marketing Association, pp.79-84. 

Bitner, Mary Jo, Faranda, William T., Hubbert, Amy R., Zeithaml, Valarie A.,(1997) “Customer 

contributions and roles in service delivery”, International Journal of Service Industry Management; 

Vol. 8 Issue 3, p193-205 

Boulding, W., Kalra A., Staelin, R., Zeithmaml, V., (1993), “A Dynamic Process Model of Service 

Quality: From Expectations to Behavioural Intentions,” Journal of Marketing Research,v30,pp.7- 

27 

Eangel, J. F., Blackwell R., (1982) Consumer Behavior New York Dryden Press. 

Gardner, Meryl Paula, (1985), “Mood States and Consumer Behavior: A Critical Review” Journal 

of Consumer Research; Vol. 12 Issue 3, p281-300 

Glass, D.C. & Singer, J.E. (1972). Urban Stress: Experiments on noise and social stressors. New 

York: Academic Press. 

Gore P, Madhavan S, McClung G, Riley DA., (1994), "Consumer Involvement in Nonprescription 

Medicine Purchase Decisions". Journal of Health Care Marketing, Vol 14(2):16-23. 

Grove, Stephen J., Fisk, Raymond P., (1997), “The impact of other customers on service 

experiences: A Critical Incidence Examination of “Getting Along”” Journal of Retailing; Vol. 73 

Issue 1, p63-87 

Gwynne, Anne L., Devlin, James F., (2000), “The Zone of Tolerance: Insights and Influences” 

Journal of Marketing Management; Vol. 16 Issue 6, p24 

Hui, Michael K., Bateson, John E.G., (1991), “Perceived Control and the Effects of Crowding and 

Consumer Choice on the Service “ Journal of Consumer Research;Vol. 18 Issue 2, p174-184 

Isen, Alice M. (1984), “Toward Understanding the Roleof Affect in Cognition”, In Handbook of 

Social Cognition, R. Wyer and T. Srull, eds. Hillsdale, NJ: Lawrence Erlbaum Associates, 179- 

236. 

Kalra, A., (1991), “An Expectations Based Model of Service Quality Assessment,” paper 

presented at the American Marketing Association Summer Educators’ Meeting, Chicago, IL. 

Knowles, Patricia A., Grove, Stephen J., Pickett, Gregory M., (1993), “Mood and the Service 

Customer” Journal of Services Marketing; Vol. 7 Issue 4, p41-52 

Machleit, Karen A., Eroglu, Sevgin A., Mantel, Susan Powell, (2000), “Perceived Retail Crowding 

and Shopping Satisfaction: What Modifies This Relationship?” Journal of Consumer Psychology; 

Vol. 9 Issue 1, p29-42 

Martin, Charles L., (1996), “Consumer-to-consumer relationships: Satisfaction with other 

consumers' public behavior” Journal of Consumer Affairs; Vol. 30 Issue 1, p146-69 

Martin, Charles L., Pranter, Charles A., (1989), Compatibility Management: Customer-To- 

Customer Relationships In Service Environments” Journal of Services Marketing; Vol. 3 Issue 3, 

p5-15 

301

McGrath, Mary Ann, Otnes, Cele, (1995), “Unacquainted influencers: When strangers interact in 

the retail setting” Journal of Business Research; Vol. 32 Issue 3, p261-272 

Mehrabian, A., & Russell, J.A. (1974). An approach to environmental psychology. M.I.T. Press, 

Cambridge, MA. 

Parker, Cathy, Ward, Philippa, (2000),”An analysis of role adoptions and scripts during customerto-customer 

encounters” European Journal of Marketing; Vol. 34 Issue 3/4, p341-458 

Peterson R. A. and M . Sauber (1983) "A Mood Scale for Survey Research" in Murphy P et al. 

(eds) American Marketing Association Educators Proceedings (409-414) Chicago: AMA. 

Schwartz, S. H., Lilach Sagiv (1995), Identifying Culture-Specifics in Content and Structures of 

Values,” Journal of Cross-Cultural Psychology, V 26 (January) 92-116 

Schwartz, S.H. (1992). Universals in the Construct and Structure of Values: Theoretical Advances 

and Empirical Tests in 20 Countries. In M.P. Zanna (Ed.). Advances in Experimental Social 

Psychology. CA: Academic Press 

Swinyard, William R.(1993), “The effects of mood, involvement, and quality of store experience 

on shopping intentions”Journal of Consumer Research; Sep93, Vol. 20 Issue 2, p271-280 

Weyant, J. M. (1978). Effects of mood states, costs, and benefits on helping. Journal of 

Personality and Social Psychology, 36, 1169-1176. 

White, R.W. (1959). Motivation reconsidered: the concept of competence. Psychological Review, 

66, 297-333. 

Zaichowsky, J. L.,(1985), “Measuring the Involvement Construct” Journal of Consumer Research, 

12, 341-352 

302

Table…1 

Estimated Marginal Means 

Factor Factor Level Marginal Means Range 

Bad 3.70 

Mood Good 2.63 1.07 

No 3.78 

Prior Expectations Yes 2.58 1.20 

No 3.38 

Control Over Activities Yes 2.97 0.41 

No 3.80 

Control Over Outcome Yes 2.55 1.25 

Conservative 3.43 

Personal Value System Open to change 2.95 0.48 

High 3.46 

Involvement Low 2.90 0.56 

303

Table…2 

Tests of Between-Subjects Effects 

Dependent Variable: Incompatibility Reaction 

Type III Sum Mean 

η 

Factor 

of Squares Df Square F Sig. 

2 (Eta 

Squared) 

Corrected Model 422.787(a) 13 32.522 83.840 .000 0.893 

Intercept 3238.513 1 3238.513 8348.651 .000 0.954 

Control over activities 13.612 1 1.612 35.092 .000 0.081 

Control over outcome 125.000 1 125.000 322.241 .000 0.546 

Expectations 115.200 1 115.200 296.977 .000 0.526 

Personal Value System 21.012 1 21.12 54.169 .000 0.119 

Mood 86.112 1 86.112 221.992 .000 0.457 

Involvement 25.313 1 25.313 65.254 .000 0.140 

Personal Value System * 

Mood 

1.012 1 1.012 2.610 .107 0.002 

Mood * Involvement .012 1 .012 .032 .858 0.000 

Expectations * Mood 1.250 1 1.250 3.222 .074 0.002 

Control over activities * 

Mood 

6.612 1 6.612 17.047 .000 0.031 

Expectations *Involvement 11.250 1 11.250 29.002 .000 0.068 

Control over outcome * 

Involvement 

16.200 1 16.200 41,762 .000 0.103 

Control over activities * 

Expectations 

,.00 1 .200 .516 .473 0.002 

Error 118.700 306 .388 

Total 3780.000 320 

Corrected Total 541.488 319 

R 2 =0.921 (Adjusted R 2 =0 .893) 

Table…3 

Signal/Noise Ratio Results 

Outer Array 

1 2 2 1 

1 2 1 2 

Inner Array 1 1 2 2 Mean SD S/N Ratio 

1 1 1 4.95 2.70 3.15 4.05 3.71 0.99 -11.5 

1 2 2 3.80 2.40 1.95 2.60 2.68 0.79 -8.8 

2 1 2 3.70 3.10 2.70 1.85 2.83 0.77 -9.5 

2 2 1 5.90 3.25 2.80 2.90 3.71 1.47 -12.0 

304

Run 

Control 

Over 

Activities 

Figure…1 

6 2 

Design Matrix for 2 − 

iv Experiment 

Control 

Over 

Outcome 

Expectations 

Personal 

Value System 

Mood Involvement 

1 Low Low No Conservative bad Low 

2 High Low No Conservative Good Low 

3 Low High No Conservative Good High 

4 High High No Conservative bad High 

5 Low Low Yes Conservative Good High 

6 High Low Yes Conservative bad High 

7 Low High Yes Conservative bad Low 

8 High High Yes Conservative Good Low 

9 Low Low No Progressive bad High 

10 High Low No Progressive Good High 

11 Low High No Progressive Good Low 

12 High High No Progressive bad Low 

13 Low Low Yes Progressive Good Low 

14 High Low Yes Progressive bad Low 

15 Low High Yes Progressive bad High 

16 High High Yes Progressive Good High 

High, Yes, Good, Progressive = + in design Matrix 

Low, No, Bad, Conservative= - in design Matrix 

305

Figure…2 (a) 

Combination of L4 Taguchi Arrays 

Outer Array 

1 2 2 1 

1 2 1 2 

Inner Array 1 1 2 2 

1 1 1 

1 2 2 

2 1 2 

2 2 1 

Figure…2 (b) 

Re-arrangement of L4 Taguchi Arrays into Design Profiles 

Ru 

n 

Control Over 

Activities 

Control Over 

Outcome 

Expectation 

s 

Personal Value 

System 

Mood Involvement 

1 

No No No Conservative Bad Low 

2 

No No No Conservative Good High 

3 

No No No Progressive Bad High 

4 No No No Progressive Good Low 

5 No Yes Yes Conservative Bad Low 

6 No Yes Yes Conservative Good High 

7 No Yes Yes Progressive Bad High 

8 No Yes Yes Progressive Good Low 

9 Yes No Yes Conservative Bad Low 

10 Yes No Yes Conservative Good High 

11 Yes No Yes Progressive Bad High 

12 Yes No Yes Progressive Good Low 

13 Yes Yes No Conservative Bad Low 

14 Yes Yes No Conservative Good High 

15 Yes Yes No Progressive Bad High 

16 Yes Yes No Progressive Good Low 

High, Yes, Good, Progressive = + in design Matrix 

Low, No, Bad, Conservative= - in design Matrix 

306

307

Quality, value, satisfaction, trust, a

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?