Conference Sessions - Jesse H. Jones Graduate School of ...

More documents

Recommendations

Info

TC04 3 - Is it a Fad or Necessity? Measuring the Effectiveness of Social Media on E-tailers Jun Yang, Assistant Professor, University of Houston Victoria, School of Business, 14000 University Blvd, Sugar Land, TX, 77479, United States of America, yangj@uhv.edu, Jungkun Park The rapid development of online communities and social networks has dramatically changed the way how marketers usually work. Consumers actively write reviews and share their experiences on social networking sites such as Facebook, Youtube and Twitter. Instead of treating consumers as passive information receivers, marketers begin to consider consumers as active co-producers of content and information. Marketing strategies such as ‘seeding’ campaigns in online communities and firms’ participation in social network sites are commonly adopted in practice. Though social media has recently become the spotlight of both practitioners and academies, there is limited research to study its impact on firms, especially on retailers. Furthermore, since the beginning of ecommerce, many studies have been conducted to identify the key determinants of successful e-tailing. However, there seems to be lack of consensus on some of the factors. This study tries to fill these research gaps. We have collected a dataset on the top 500 e-tailers. The dataset contains detailed information on the features provided by each e-tailer, and its financial performance. We seek to answer the following research questions. First, we empirically identify the determinants for a successful e-tailer, since there is still lack consensus in the literature. Second, after controlling those determinants, we would like to measure the influence from social media on those e-tailers’ performances. Furthermore, we want to test whether those influences will differ across various channel structure types and across different product categories. Empirical results and managerial implications are included. ■ TC04 Legends Ballroom V Econometric Methods I: General Contributed Session Chair: Sridhar Narayanan, Stanford University, 518 Memorial Way, Stanford, CA, 95014, United States of America, sridhar.narayanan@stanford.edu 1 - A Cigarette, a Six Pack or Porn? The Complementarity of Vices Rachel Shacham, Carlson School of Management, University of Minnesota, 321 19th Avenue South, Suite 3-150, Minneapolis, MN, 55401, United States of America, rshacham@umn.edu, Peter Golder, Tulin Erdem Using statistical copulas, we develop an empirical model that allows us to study the levels of complementarity between different vices while accounting for self-selection. The method developed in this study is particularly well suited to the issues that occur while studying vice (in contrast to other types of goods). In particular, we allow the level of complementarity to differ between addicts and other users. We estimate the model using a unique dataset that contains detailed individual-level data over time. 2 - Handling Endogenous Regressors by Joint Estimation using Copulas Sungho Park, Assistant Professor, Arizona State University, P.O. Box 874106, Tempe, AZ, 85287, United States of America, spark104@asu.edu, Sachin Gupta We propose a method to tackle the endogeneity problem without using instrumental variables. The proposed method models the joint distribution of the endogenous regressor and the error term in the structural equation of interest (the structural error) using a copula method, and makes inferences on the model parameters by maximizing the likelihood derived from the joint distribution. Using a series of simulation studies and an empirical example, we show that the proposed method captures the dependence structure between the endogeneous regressor and the structural error well enough to overcome the endogeneity problem. Other properties of the proposed method are discussed. 3 - Improving Predictive Validation Steve Shugan, University of Florida, 2030 NW 24th Avenue, Gainesville, FL, 32605, United States of America, sms@ufl.edu Predictive validation is very popular in management science and marketing science for many reasons including the scientific principle that valid theories should make valid predictions. However, when predictive validation and testing use the same data, it has been shown that invalid theories can imply incorrectly specified models that defeat predictive validation. Moreover, wrong models can out-predict the true model (that represent the actual data generating process), both in-sample and out-ofsample. Consequently, focusing only on fit or predictive validation can result in wrong implications. For example, as shown in the paper, better predicting response functions often underestimate optimal expenditures and imply wrong strategies. This paper shows (analytically and through simulation) that combining several predictive metrics allows detection of wrong response functions better than all popular extant metrics (absolute-mean-error, square-error, likelihood, and so on). The analytical proofs show how to improve predictive validation when those wrong response functions employ dampening. Dampening is a special type of bias that (as shown in the paper) allows wrong response functions to out-predict the true response function that represents that actual response to the decision variables (e.g., price, advertising, etc.). The simulations show how to improve predictive validation in the general case. The paper also discusses the interpretation of predictive validation in terms of MARKETING SCIENCE CONFERENCE – 2011 20 inferential statistics, a Bayesian decision perspective, the classical scientific method, and the modern machine learning theory perspective. The paper is available from the author. 4 - Regression Discontinuity with Unobserved Score Sridhar Narayanan, Stanford University, 518 Memorial Way, Stanford, CA, 95014, United States of America, sridhar.narayanan@stanford.edu, Kirthi Kalyanam Regression Discontinuity (RD) designs estimate treatment effects in situations where the treatment is based on an underlying continuous ‘score’ variable, with treatment taking place if the score crosses a discete threshold and not otherwise. The local average treatment effect is measured by comparing outcomes for subjects whose scores are just above and just below the threshold, with the latter serving as a control group for the former. We extend the scope of RD to contexts where the score or the threshold are not fully observed, but only components of the score, or covariates that explain treatment are observed. The method involves estimating scores using a first stage empirical approximation, which involves fitting a binary choice model for treatment as a function of observed score components or other covariates. Next, the outcomes for individuals with estimated score just above the threshold are compared with those just below the threshold to obtain the treatment effect, as in a standard RD approach. Since a binary choice model has an inbuilt threshold of zero, this approach works even when the actual threshold on the true score function is unobserved. We show analytically the conditions under which the method uncovers a valid treatment effect. To demonstrate the methodology, we use a casino marketing setting where the exact scores used by the casino to decide on the treatment (offers mailed to consumers) are observed. We use a Monte Carlo simulation to establish the empirical conditions required to estimate the treatment effect. Finally we show an application to pharmaceutical detailing, where the scores are unobserved. The estimates using our our proposed approach generate new insights that add to the literature. ■ TC05 Legends Ballroom VI New Product III: Adoption Contributed Session Chair: Mark Ratchford, Assistant Professor of Marketing, Vanderbilt University, Owen Graduate School of Management, 401 21st Avenue South, Nashville, TN, 37203, United States of America, Mark.Ratchford@owen.vanderbilt.edu 1 - An Investigation of Scales for Consumer Innovativeness Masataka Yamada, Professor, Nagoya University of Commerce and Business, 4-4 Sagamine, Komenoki-cho, Nissin-shi, 470-0193, Japan, myamada@nucba.ac.jp, Toshihiko Nagaoka Last year at Cologne, we reconsidered not only the consumer innovativeness but also the theory of innovation diffusion itself focusing on the construct of consumer innovativeness. We proposed a new way of viewing the diffusion of innovation using Carnap’s theoretical construct and disposition concept. We introduced an intermediate level of abstraction construct between theoretical construct and disposition concept. We name it T-D mixture. This study reviews the measurement scales of each level of innovativeness. Then we reposition some of the scales as the scales for the T-D mixture. We can expect better predictions of actual adoption behavior by these scales. 2 - A Multivariate Analysis of Pre-acquisition Drivers of Technology Adoption Mark Ratchford, Assistant Professor of Marketing, Vanderbilt University, Owen Graduate School of Management, 401 21st Avenue South, Nashville, TN, 37203, United States of America, Mark.Ratchford@owen.vanderbilt.edu, Jeffrey Dotson This study develops and empirically tests a new parsimonious multiple-item scale to measure consumers’ propensities to adopt new technologies. We show that a consumer’s likelihood to embrace new technologies can reliably be measured by a 14item index that combines assessments of consumers’ positive and negative attitudes towards technology. Consistent with prior work on technology readiness, we show four distinct dimensions of consumers’ technology adoption propensity: two inhibiting factors, dependence and vulnerability, and two contributing factors, optimism and proficiency. We develop the index on a cross-sectional dataset of U.S. consumers then establish the validity of each of the four component scales on two dissimilar validation sets. Next, using a multivariate probit model, respondents’ scores on the resulting index sub-scales are combined with demographic factors to examine the antecedents of adoption of a range of technologies. Distinct segments of adopters emerge. The results are of interest to practitioners interested in learning about the drivers of technology adoption for new products.
■ TC06 Legends Ballroom VII Competition III: General Contributed Session Chair: Vincent Mak, Judge Business School, University of Cambridge, Judge Business School, Trumpington Street, Cambridge, CB2 1AG, United Kingdom, v.mak@jbs.cam.ac.uk 1 - What if Marketers Put Customers Ahead of Profits? Scott Shriver, Stanford GSB, 518 Memorial Way, Stanford, CA, United States of America, scott.shriver@gsb.stanford.edu, V. “Seenu” Srinivasan We examine a duopoly where one of the firms does not maximize profit, but instead maximizes customer surplus subject to a profit constraint. The model assumes customer willingness to pay for quality is uniformly distributed and that customers follow a simple decision rule: when presented with two products of known quality and price, purchase one unit of the product which maximizes surplus, or if surplus is negative for both products, elect not to purchase any product. We further assume that firms’ marginal cost of production is convex (quadratic) in quality. Competition between firms is modeled as a two-stage game, which is solvable by backward induction. In the first stage, firms choose product quality levels sequentially, fully anticipating subsequent price competition. In the second stage, firms take qualities as given and choose prices simultaneously in accordance with a Nash equilibrium. Two possibilities are considered: (a) the first mover is the profit maximizing firm, and (b) the first mover is the surplus maximizing firm. We compare the results to the corresponding base case of Moorthy (1988), where both firms are profit maximizing. We find that firms can deliver significant additional value to their customers by forgoing small amounts of profit. The effectiveness of this strategy depends upon which firm is the first mover. In the case that the surplus-maximizing firm moves first, 5%-10% increases in customer surplus “cost” 10%-20% of potential profits. By contrast, when the profit-maximizing firm chooses quality first, we find that sacrificing 20% of profits is sufficient for the surplus-maximizing firm to more than triple the customer surplus it would have provided under a profit-maximizing objective. 2 - Gaining from Imitative Entry: Dynamic Durable Pricing with Rational Consumer Expectations Lu Qiang, City University of Hong Kong, ACAD P7701, 83 Tat Chee Avenue, Kowloon Tong,Hong Kong - PRC, qianglu@cityu.edu.hk, Wei-yu Kevin Chiang This paper investigates the impact of imitative entry on intertemporal pricing strategy of an innovator (brand-name company) selling a new durable in a two-period dynamic framework. While acting as a monopolist in the first period, the innovator faces competitive entry of an imitator in the second one. The market consists of loyal versus non-loyal segments of consumers who are heterogeneous in their valuation of product. The loyal customers buy only from the innovator. Before deciding when to buy, they rationally anticipate the pricing policy to be followed by the innovator who is unable to credibly commit in advance to the second-period price. On the other hand, the non-loyal customers are price-sensitive and only act in the second period. In making a purchase decision, they compare the benefit of buying from the innovator against that from the imitator. Contrary to the conventional wisdom, the equilibrium of a two-stage pricing game between the innovator and the imitator indicates that imitative entry may potentially benefit the innovator. We provide specific explanations and implications to this counter-intuitive finding by looking further into the dynamic pricing strategy of the innovator and the corresponding purchasing choice of the rational consumers. 3 - KFC and McDonald’s Entry in China: Competitors or Companions? Qiaowei Shen, Assistant Professor of Marketing, The Wharton School, University of Pennsylvania, 3730 Walnut Street, JMHH 700, Philadelphia, PA, 19104, United States of America, qshen@wharton.upenn.edu, Ping Xiao In this paper, we study the entry of McDonald’s and KFC, the two largest fast food chain restaurants in the world, in China, which is the world’s largest emerging market. Our data covers the entire history of entry by the two chains from 1987 when the first KFC outlet opened in Beijing up to year 2007, when more than 200 cities have KFC or McDonald’s or both. We are particularly interested in how the duopoly western fast food chains may influence each other in their decision to enter a new city or open additional outlets in an existing market (city). We find that the scale of rival presence as indicated by the cumulative number of outlets has a positive effect on a chain’s entry decision after controlling for local economic conditions, regional time-invariant unobservables and time trend. We find evidence that such positive effect is due to firm learning about uncertain market demand. Furthermore, the effect varies by time period (e.g. before 1999 vs. after 1999) and by city type (big city vs. small city). MARKETING SCIENCE CONFERENCE – 2011 TC07 21 4 - Dominance and Innovation in a Dynamic Macro Environment Vincent Mak, Judge Business School, University of Cambridge, Judge Business School, Trumpington Street, Cambridge, CB2 1AG, United Kingdom, v.mak@jbs.cam.ac.uk, Jaideep Prabhu, Rajesh Chandy, Chander Velu One of the most widely contested issues in innovation research is whether dominant firms tend to be lethargic or pioneering with innovations. Both theoretical and empirical evidence is divided on this topic. In this paper, we propose one way to reconcile these differences. We build a game-theoretic model that highlights a key driver of innovation by dominant as well as less dominant firms: the macro environment faced by these firms. Specifically, the model focuses on how changes in market profitability due to exogenous changes in competing firms’ macro environment affect the innovation strategies of these firms. We show that in environments with rapidly declining profitability, dominant firms are likely to be lethargic and innovate incrementally, while in environments with rapidly growing profitability, dominant firms innovate radically and ahead of competitors. Surprisingly, we also find that a firm in an environment with rapidly growing profitability might innovate at a later time (if at all) and earn a lower expected profit when its probability of succeeding with a radical innovation increases. Finally, we examine a case in which the macro environment is profitable throughout, and find that in such situations a hybrid innovation pattern can emerge whereby the dominant firm innovates radically but late. ■ TC07 Founders I ASA Special Session on the Marketing-Statistics Interface - III Cluster: Special Sessions Invited Session Chair: Michael Braun, Massacusetts Institute of Technology, 100 Main Street, Cambridge, MA, United States of America, braunm@mit.edu 1 - Customer Waiting Time and Purchasing Behavior: An Empirical Study of Supermarket Queues Andrés Musalem, Duke University, Fuqua School of Business, 100 Fuqua Drive, Durham, NC, 27708, United States of America, andres.musalem@duke.edu, Yina Lu, Marcelo Olivares, Ariel Schilkrut The design of services requires making decisions about service features, employee selection and training, staffing levels and standards of service quality. The traditional approach has been to design and manage services to attain a quantifiable service standard – for example, call centers are designed to guarantee that a fraction of their customers (say 95%) are served within a given waiting time (e.g., 30 seconds). These design decisions usually involve trade-offs between the costs of sustaining a certain service standard versus the value that customers attach to this level of service. In contrast with extant literature that focuses on subjective measures of service quality, in this paper we measure the effect of objective measures of service quality – in particular, waiting time of customers in a queue – on actual customer purchasing behavior. Accordingly we develop a methodology which can be used to attach a financial tag on the implications of having customers waiting for service. Our paper uses a novel data set which was collected through digital cameras and image recognition software. Conducting a pilot study in the fresh deli section of a large supermarket, we collected data on queue lengths and number of staff members providing service every 30 minutes during a 6 month period. In addition, we obtained point-of-sales (POS) data from all relevant transactions during this time period. The time-stamp of the POS transaction is used to link the snapshot data from the queue with the customer purchase data and data augmentation methods are used to account for the uncertainty about the exact queue length that customers experience. A key feature of the supermarket we study is that a large fraction of their purchases are from loyalty card holders. This is used to construct a panel data set of customer purchases, which enables us to control for customer heterogeneity in purchasing behavior using Bayesian methods. 2 - Forecasting Customer Purchase Rates Incorporating Temporal Variation Luo Lu, Stanford University, Sequoia Hall, Stanford, CA, United States of America, luolu@stanford.edu, Zainab Jamal When predicting the future purchasing patterns of customers, the expected number of transactions in a future period is an important factor in the “lifetime value” calculations for each individual customer. Among the extant models that provide such capabilities, the BG/NBD model stands out. Under this framework, a major assumption is that the number of transactions made by a customer follows a Poisson process with a fixed transaction rate. However, in the real world the transaction rates may vary based on time. For example, in our empirical data from an online digital sharing service, we found a “seasonal pattern” which indicates that the customers intend to make more purchases in summer and winter. We also found that the customers make less purchases as time goes by, indicating an “annual pattern”. We replaced the homogeneous Poisson process in the BG/NBD model by a nonhomogeneous one to include the seasonal and yearly factors in the transaction rates and used a Bayesian framework to estimate the model at an individual level. The results demonstrate the effectiveness of the proposed methods.
Page 1 and 2: How to Navigate the Contributed Ses
Page 3 and 4: address a target group. i.e., shoul
Page 5 and 6: ole that prior distributions play i
Page 7 and 8: 4 - Understanding Customers` Substi
Page 9 and 10: shift in the demand curve. Most of
Page 11 and 12: marketers may predict future custom
Page 13 and 14: ■ TB06 Legends Ballroom VII Compe
Page 15 and 16: 3 - The Impact of Free-trial Promot
Page 17 and 18: 3 - Cross-cultural Differences in B
Page 19: of the sequence of evaluation, a pa
Page 23 and 24: and on the transaction utility they
Page 25 and 26: een developed would help marketers
Page 27 and 28: Thursday, 3:30pm - 5:00pm ■ TD01
Page 29 and 30: interest. To evaluate the performan
Page 31 and 32: 2 - What You Don’t Know Can’t H
Page 33 and 34: ■ TD12 Champions Center II Biddin
Page 35 and 36: while maintaining market growth rat
Page 37 and 38: knowledge. However, buyers with a h
Page 39 and 40: ■ FA07 Founders I Panel Session:
Page 41 and 42: ecoming more acceptable, the dilemm
Page 43 and 44: 2 - Choosing the Right Plan? Asymme
Page 45 and 46: 4 - A Framework for Unifying Differ
Page 47 and 48: ■ FB06 Legends Ballroom VII Chann
Page 49 and 50: 3 - Money-back Guarantees: The Grea
Page 51 and 52: 2 - The Impact of Consulting on Buy
Page 53 and 54: 2 - Data or Structure? Using a Fiel
Page 55 and 56: 4 - Green Lifestyle Adoption: Shopp
Page 57 and 58: 2 - Determining Consumers’ Discou
Page 59 and 60: 3 - Investigating Salespeople Turno
Page 61 and 62: 2 - One-stop Shopping: A Double Edg
Page 63 and 64: 3 - Which Link to Click - Sponsored
Page 65 and 66: valuation of the high-end consumers
Page 67 and 68: 2 - A Machine Learning Approach to
Page 69 and 70: 4 - Social Ties and User Generated
Page 71 and 72:
■ SA02 Legends Ballroom II Twitte
Page 73 and 74:
■ SA05 Legends Ballroom VI Game T
Page 75 and 76:
differ depending on their respectiv
Page 77 and 78:
2 - The Impact of Online Referrals
Page 79 and 80:
Saturday, 10:30am - 12:00pm ■ SB0
Page 81 and 82:
2 - R&D Spillover and Product Diffe
Page 83 and 84:
2 - The Advertising Role of Profess
Page 85 and 86:
4- Optimal Allocation of Marketing
Page 87 and 88:
3 - Information Communities: The Ne
Page 89 and 90:
■ SC06 Legends Ballroom VII Consu
Page 91 and 92:
and context is unclear. The current
Page 93 and 94:
A Agarwal, Nipun SB14 Alexandrov, A
Page 95 and 96:
F Fan, Tingting FC10 Fang, Eric TA0
Page 97 and 98:
Schweidel, David TC07, FA02, FB02,
Page 99:
Friday, 1:30pm - 3:00pm FC01 Choice
show all

Conference Sessions - Jesse H. Jones Graduate School of ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?