AT Internet

www.atinternet.com 

Online Intelligence Solutions 

Leveraging Web Analytics 

to 

Optimize Advertising Revenue 

Visitor Segmentation as a Critical Capability to Ad Revenue Optimization 

Jean-Marie Camiade, Knowledge manager 

AT Internet White paper


Table of Contents 

Introduction.........................................................................................................................................................2-3 

What do Online Advertisers Look For in a WebSite?...................................................................................4 

The Keys to Success............................................................................................................................................4-5 

Know your own traffi c, and know it well.....................................................................................................6-8 

Optimizing Advertising Revenue......................................................................................................................9 

Increasing Inventory..................................................................................................................................9 

Increasing value of inventory........................................................................................................10-11 

Increasing percentage of sold inventory...................................................................................12-13 

Ideally, we would like to get into people’s mind, but how?..................................................................14 

Rich Media Ads: What works best for what audience?.....................................................................14-16 

Next Step..................................................................................................................................................................18 

About AT Internet.................................................................................................................................................19 

Contact.....................................................................................................................................................................20 


Media 

AT Internet White paper



Introduction 

The Web will be the hottest tactic among advertisers and marketers in 2009 according 

to executives surveyed in January 09 by Reardon Smith Whittaker. Asked which 

tactics most interested the marketers themselves, nearly seven out of 10 said online 

marketing. In this economic downturn, it is defi nitively good news to hear if you’re a 

publisher using advertising to generate revenue. 

Averaged realized CPMs on ad networks 

ranged from $0.60 to $1.10, versus $10 

to $20 in direct sold display inventory, or 

only 6 to 11% of direct pricing. 

The following graphs represent aggregated 

data for 7 large US publishers (media 

companies) in 2007 and provide a good 

indication of the trends between Direct 

sales and Intermediary sales. 

However, a 2008 benchmark study conducted 

by the IAB and Bain & Company suggested 

that online publishers were increasingly 

turning to sales intermediaries 

known as ad networks to sell off excess 

inventories. The use of ad networks surged 

from 5% of total ad impressions sold in 

2006 to 30% in 2007 among branded/large 

publishers. Overall, for large and small publishers, 

unverifi ed studies were even claiming 

that up to 80% of all ad space went 

unsold by website directly; unsold ad space 

was sent to ad networks to monetize. 

Media 

- 2 - 

White paper AT Internet



So what to do? Publishers need to better support the value of their inventory. Instead 

of selling off more and more ad space to intermediaries at 6 to 11% of what it is worth 

when sold directly, Publishers should start leveraging their Web Analytics investment 

assetto its full potential. Most businesses look at their Web Analytics solution just as 

“General Web Traffi c Analysis” (how many visitors, how many page views, how much 

time spent on site, etc). 

Businesses must realize that the use of 

web analytics in this way can be a very 

powerful tool which delivers signifi cant 

business advantage. 

The rest of this document will hopefully 

provide you with a new appreciation of the use of analytics when it comes to optimizing 

advertising revenue on a publisher’s web site. 

Media 

- 3 - 




What do Online Advertisers Look For in a WebSite? 

The rest of this document will hopefully provide you with a new appreciation of the 

use of analytics when it comes to optimizing advertising revenue on a publisher’s web 

site. 

● Question: 

What factors do you look for when selecting the web sites where you will buy advertising? 

● Some of the answers: 

“...the number of visitors/hits/etc is pretty meaningless per se - It’s a question of a match 

with the audience.” 

“...without a doubt the best results have occurred when the ads were on a site with an audience 

that was a perfect fi t for our product”. 

“... audience matching our target niche”. 

The Keys to Success 

Audience Segmentation. Behavioral Analysis. Targeting capabilities. That pretty much 

sums it up. To outperform as a media publisher, you have to understand what an advertiser 

is looking for. And all that an advertiser is looking for is to reach his target 

audience. You might think that there is more to it? Not really. It’s that simple. All year 

long, advertisers (his internal staff and/or his advertising agency) try to answer the 

following questions in this order: 

● What is my target audience? 

Even though it’s ultimately the advertiser’s role to fi gure out what his target audience 

is (market research, customer database, surveys, focus group, etc), you as a publisher 

can help your client advertisers refi ne and even test new target audience segments 

that could have been overlooked, and that could end up being very successful for 

them. Your Ad Sales Team needs to get into a partner mentality with their advertiser 

clients. In a competitive world, diff erentiating yourself from other publishers by bringing 

additional value to your clients 

is always a good strategy. 

Media 

- 4 - 




● Where is my target audience? 

Using Analyzer NX and its powerful segmentation capabilities, you will be able to know 

with great level of details which and how many of your advertiser’s audience can be 

reached on your web site. 

● How to best reach my target audience? 

Leveraging the same powerful segmentation capabilities from Analyzer NX , it becomes 

quite simple to understand what audience type does what on your web site, thus allowing 

your advertisers to reach their target audience with great level of effi ciency. 

● How to optimize my investment of reaching my target audience? 

Optimizing marketing branding or marketing acquisition investment through advertising 

is directly related to the level of details and confi dence your advertisers has in the 

publisher (you) to answer questions such as how many, where and how to reach the 

right target audience. There are many ways to reach a target audience, and it doesn’t 

always have to be advertising on premium sections (like home pages) which could be 

expensive (or often sold out), and that could be an overkill for many advertisers with 

lower marketing budgets. 

Media 

- 5 - 




Know your own tra c, and know it well 

Before anything else, it is essential that you get to know your own traffi c with 2 objectives 

in mind. You cannot be in control of your advertising success if you haven’t done 

this critical work fi rst. All aspects of the enterprise are impacted by it: your editorial 

staff , your ad sales team, even your own traffi c acquisition eff orts. There are two main 

tactics, two diff erent approaches to understanding your own traffi c: 

Tactic 1: From Advertiser to Audience. As a publisher, you certainly know who are the 

largest advertisers in the market place that you would like to go after, whether to win 

their business, or to increase their marketing budget with you. Each of these advertisers 

goes after a specifi c audience (demographic, technographic, psychographic, interest 

based, etc), so the objective here is to fi gure out how many, where and what 

each of these “specifi c audiences” do on your web site. Once you know the answers, it 

becomes easy to package it and present it to the advertisers with a compelling story 

and marketing plan. 

How to get the answers? 

The objective is to segment your visitors (as many as possible, as long as it represents a 

statically relevant sample of your audience) based on the advertisers’ audience known 

criteria. Example: if advertiser A is looking to reach “female, age 20 to 35, practicing 

sports”, then you need to fi nd a way to capture that information and log it providing 

you with invaluable web analytics segmentation criteria. The information can be captured 

during registration or subscription processes you might already have on your 

web site. If not, then you can run a survey and get enough people to respond to it and 

answer key demographic and interest questions that will support the identifi cation of 

your advertisers’ audience. 

Media 

- 6 - 




Why permanent? 

Because you don’t want to have to recapture the same information over and over 

again. Visitors will most likely answer a survey just a couple of times, register only once, 

etc. But when they come back (and for each successive visit), we still want to track and 

analyze their behavior as part of the appropriate audience segment. Thus, we need to 

make sure that when people do respond to a survey or register for example, that all the 

relevant and important information are being captured and associated permanently 

with the visitor (using permanent cookies). Once the data collection is completed, you 

can then confi gure the Data using “Categories of visitors” and slice and dice all the web 

traffi c data by visitor segments (see screen shot above). All of a sudden, your web traffi 

c data takes on a whole new dimension and can deliver signifi cantly improved value. 

As the next screen shot suggests, the top 10 advertisers usually represent 70% of online 

ad revenue (with the top 25 representing 81%, and the top 50 representing 90% 

of the online ad market). So concentrate on the top advertisers for your market. If needed, 

interview them (or their advertising agency) and design a segmentation strategy 

with these advertiser’s target audience in mind. 

Media 

- 7 - 




Source: Internet Advertising Bureau & PricewaterhouseCoopers - October 2008 

Tactic 2: From Audience to Advertiser. In this scenario, you want to get as much audience 

profi le information as possible, and design your advertising media plan (advertising 

companies to go after, sales pitch, media kit, etc) according to the diff erent 

segments you’ll be able to uncover from your web site audience. For example, if you 

discover that many of your audience have pets (cats/dogs/etc), then it makes sense to 

contact animal food supply manufacturers to advertise on your web site. 

How to get the answers? 

Exactly the same way as with tactic 1. List all the known segment criteria that are 

important for advertisers (gender, age, income, interest, industry, job, pet owner, car 

model they drive, etc) and develop a data capture plan that fi ts with your web site activities, 

and that doesn’t negatively impact the user experience. All data doesn’t need 

to be captured at once. The most important ones fi rst (like age and gender), and then 

as time goes by, you can start collecting more granular information to make your segments’ 

profi le richer. Information can be captured during a registration or subscription 

form (or on the thank you page if you think that requesting too much information during 

the process might negatively impact your conversion rate), using a survey, game 

contests, poll questions, etc. The important thing to keep in mind is that you always 

need to collect a statically sound sample for each segmentation criteria, and that these 

data criteria must be captured as permanent segmentation criteria as previously explained. 

Media 

- 8 - 




Optimizing Advertising Revenue 

Increasing Inventory 

There are 3 key performance indicators that contribute directly to the positive or negative 

growth of your advertising inventory’s worth. 

● Increasing Unique Visitors: Where do NEW visitors come from? Search Engines, banner 

campaigns, sponsorship placement, online partners, social media sites, others? How 

about returning visitors? It’s just a matter of making sure all your traffi c acquisition 

tactics work and produce the results you expect. Always keep an eye on the ROI 

though; as increasing unique visitors cannot be done at any cost, which is why social 

media marketing and word of mouth (WoM) is so vital to today’s online marketing 

departments. 

● Increasing the number of Visits per Visitor: Acquiring new visitors is one thing, but 

getting them to come back is another. We live in a world where information is everywhere. 

It’s overwhelming. We don’t have the time anymore to browse around. We 

need to get to the essential. Getting to the essential is pretty much getting the right 

information at the right time. That’s why email alerts and RSS feed subscriptions have 

become more popular over the recent years (as of October 2008, the RSS adoption 

rate is at 11% according to Forester Research). So to get people back to your web site, 

optimizing email alerts and RSS feed subscriptions are two industry best practices 

among others that should be promoted and measured constantly. In addition, the 

more people subscribe, the more chances you have to collect anonymous profi le 

information that you can turn into segmentation criteria for improved audience analysis. 

● Increasing Page Views per Visit: Finally, the last piece of advertising inventory growth is 

to keep the visitor interested in the content, and getting him to browse more pages 

per single visit. Two examples of best practices to achieving that goal are (1) to focus 

on a landing page with high traffi c and high single access ratio (or high bounce 

rate ratio), and (2) maximizing articles, stories, editorial content across several pages 

(e.g: instead of one long story in 1 single page, the content is distributed across 2 

or 3 pages, generating 2 to 3 times more advertising inventory). A side advantage 

for using such a tactic is that it helps measure the interest in the story. If the story 

requires the reader to look at 3 pages to fi nish it, then you know that if the average 

depth for that story is 1.2, then most likely the story didn’t retain the reader’s attention). 

Media 

- 9 - 




Increasing value of inventory 

To increase the value of your inventory, you simply need to give reason for why the 

price you’re charging is worth your advertiser’s money. The more legitimate data you 

can provide that shows the matching degree between your advertiser’s goals and your 

web site advertising off ering, the easier it will be for you to justify the cost of reaching 

their audience, and get the advertisers to pay for it. 

● It’s all about Audience Targeting: Advertisers are interested in reaching specifi c audiences, 

and if they can, they are willing to pay a premium price for it. Media Publishers 

must provide solid and relevant data to advertisers guaranteeing them 

success reaching their audience on your website. Such guaranty (back by numbers, 

research and analysis) justifi es the premium price you can charge. You can provide 

added value compared to other publishers who haven’t developed an in depth audience 

discovery and segmentation strategy. 

● Diff erent Advertising Programs for diff erent advertising purposes: Branding and Traffi c 

acquisition are two of the main reasons advertisers purchase online advertising real 

estate. As a publisher, you need your ad sales team to rapidly understand what your 

advertiser’s objective is so that you can recommend the best media buy strategy. 

Media 

- 10 - 




● Branding objective: if the advertiser is interested in branding, then most likely 

a CPM (cost per thousand impressions) is best for everyone. Because the objective 

is to expose someone to a message, and make sure the message is 

being seen and recorded in the visitor’s mind, usually such messages makes 

use of rich media advertising (fl ash, video, etc). Because rich media advertising 

requires a longer attention span than traditional banner ads, the best 

placements for such ads are either (1) landing pages, or (2) pages with higher 

average time spent. Landing pages because as you arrive on a new web site, 

it takes some times for the eyes to search for things, and thus a “branding” 

message has a higher chance to grab someone’s attention. Higher average 

time spent on pages because the more time someone spend on a page, the 

more likely the rich media advertisement has a chance to grab the visitor’s 

attention. Remember though that you must understand these metrics (both 

landing pages and time spent on pages) on a per audience segment basis. 

Landing pages and time spent on pages will vary from audience segment to 

audience segment (e.g: a young adult will spend their time on your site diff erently 

than a senior person for example, or man versus women). 

● Traffi c acquisition objective: if your client is interested in acquiring traffi c to 

their web site, then a CPC (cost per click) is best for everyone. Best placements 

for CPC ads are: 

- On most popular exit pages (as long as these pages are legitimate exit 

pages), 

- Bottom of pages - once someone’s fi nished reading an article, a story, 

etc. Or, after at the end of a registration process, or once someone’s 

fi nished answering a question (or a poll). The idea is pretty much to 

place CPC ads wherever a visitor has completed an action (reading, registering, 

answering, etc) and to propose the visitor with a new action: 

“here is what you should do” (like clicking on this ad). 

- Contextual ads - Everyone loves relevancy and timing. How to get the 

right information at the right time. Which is why as someone reads 

an article or a story, placing contextual ads in it, or next to it produce 

good results because they are relevant and timely (ex: the article is 

about natural animal food, and the contextual ad links people to an 

“natural animal food” advertiser). 

Media 

- 11 - 




Increasing percentage of sold inventory 

After “Increasing Advertising Inventory”, and “Increasing Inventory Value”, the fi nal step 

is increasing the percentage of sold inventory (or decreasing the percentage of unsold 

inventory). Right away, I can hear “We don’t have unsold inventory, so how could we 

reduce it”. It’s true that most of the publishers have 0% of unsold inventory and the 

reason is because everything that is not sold directly is sold through intermediaries 

such as advertising networks like Advertising.com, Revenue Science, Google Adsense 

or Yahoo! Publisher Network. As mentioned earlier in this document, it is known in 

the industry that up to 80% of all ad space goes unsold by websites directly, and that 

unsold ad space is sent to ad networks to monetize at a much lower rate than if being 

sold directly. So the objective here is not to reduce the unsold inventory per se, but 

to reduce the inventory sold through intermediaries, and maximize its CPM via direct 

sales. 

● Affi nity Content is essential: We have explained that visitor segmentation is key to 

optimizing advertising sales. It is the most eff ective way to reach a particular audience. 

However, and because audience segmentation is either poor or nonexistent, 

advertisers request their ad to be placed in a specifi c content area. For example, if the 

advertiser is a sport manufacturer, and if the publisher has a “sport category”, then 

the advertisers will want its ad to be displayed in the sport category. Why? Because 

that is where he is most likely to reach his target audience (sound familiar!). Unfortunately, 

if the sport category is sold out, the advertiser is out of luck and may decide to 

not to buy at all or wait for the category ad inventory to become available. The solution 

for the publisher is to provide the advertiser with “content affi nity” suggestions. 

Meaning, what are the top other categories/pages that visitors to the sport category 

look at as well? As a publisher, you need to develop an affi nity map of your web 

site, starting with the most popular sold out and requested content area. Once the 

affi nity map is developed, then it becomes easy for the ad sales team to propose to 

advertisers other ad placements that would reach if not 100% of the same audience, 

maybe 50%. If the price is right (meaning proportionally discounted from the sport 

category), your ad sales team should be able to sell directly ad spaces at a decent 

CPM or CPC rate that would otherwise be sold off via intermediaries at a signifi cantly 

lower rate. 

Media 

- 12 - 




● Below the fold placement: Premium placement is usually for ad spaces above the fold 

(part of the web page that doesn’t require any scrolling to be seen). Advertisers are 

usually seeking placement above the fold, and will strongly negotiate down the 

price for inventory that is below the fold, and even more for inventory that is at the 

very bottom of the page. If you have a diff erent CPM rate between above and below 

the fold, then it’s important that you measure the scrolling index, starting with web 

pages that belong to popular content areas (and usually sold out). By measuring the 

scrolling rate of high revenue potential pages, you can develop your pricing according 

to the data you have. If you fi nd out that 90% of visitors scroll all the way down 

a specifi c page, then your CPM rate should take such ratio into account. 

● Provide the ability to Test-Drive your off ering: Especially for new skeptical/hard to win 

over advertisers, look within your unsold inventory sections/pages that match your 

advertisers’ audience, and you should be able to sell some of this inventory as a test 

drive and get him comfortable with advertising on your web site. 

● Still unsold inventory? Outsource the remaining unsold inventory to the best Ad 

Network to maximize revenue. 

Media 

- 13 - 




Ideally, we would like to get into people’s mind, 

but how? 

Simple. Search keywords. It is still astonishing how few organizations leverage keyword 

searches. By monitoring one’s behavior on a web site and looking at what content/ 

pages they look at we can we imply intent and interest. And even though web analytics 

can provide us some answers, it will never be as powerful as “asking someone” 

(What are you looking for?). The search box on your web site is like a Customer Service 

representative asking “What are you looking for”, and by entering a search phrase, the 

visitor is giving you a direct access to his mind. This is ten times more powerful than 

any deep navigation path analysis you might undertake which at best can only “guess” 

what people are trying to achieve. 

External and Internal keyword search analyses are essential for example to: 

● Guidie editorial content (do you have relevant and fresh content for the top keywords 

people type on your own web site?). If not editorial, how about marketing content? 

● Driving Search Engine Marketing. If people search and fi nd good content on your 

web site, you should make sure everyone outside your site searching for the same 

keywords get to fi nd you. 

● Create new profi ling criteria than can be used to enrich or create new audience segments. 

Rich Media Ads: What works best for what audience? 

In the pursuit of a better “Advertiser/Publisher” partnership leading to increased business 

for all parties involved, you should develop your own benchmark as to what 

works and what doesn’t by audience types. For example, are men more likely to watch 

a video than women? Are 15 to 24 year old people more likely to interact with a rich 

media ad than 25 to 34 years old? This information is extremely valuable when it comes 

to recommend and sell ad spaces on your web site. 

The following screen shots are provided as KPI examples and industry benchmark information 

(US Advertisers). Top KPI Recap for Media Sites (Rich Media Example). 

Media 

- 14 - 




Click-Through Rate 

Consider Larger Creative Sizes, in-Page video for higher Click-Through Rate 

Interaction Rate 

Consider In-Page video and larger creative sizes for higher interaction rate 

Interaction Time 

Average interaction time is a measure of how long viewers are interacting with an ad 

unit. 

Your objectives are to build your own benchmark for your own market and for your 

main audience segments. Here are 3 examples of what your web analytics program 

should help you develop when it comes to adding value to the “Advertiser-Publisher” 

relationship: 

● Example 1: 

Recommend “in-page” video advertising to reach male 15 to 24 year old (if that’s the 

case on your web site) for lead generation purposes (CPC model) 

Media 

- 15 - 





Recommend rich media medium rectangle (300x250) to acquire female with yearly 

income over $60K. 


Recommend Skyscrapers (120x600) to Financial Services companies if their objective 

is branding. Why? Because that’s what generates the greatest interaction time on 

your site, allowing for better branding message communication. 

Media 

- 16 - 




Next Step 

The Media Opportunity Calculator: The Media Opportunity Calculator is a special 

“What if Scenario” dashboard that helps benchmark and simulate where the media 

company should prioritize its eff orts in order to maximize revenue and profi ts. The 

“What if” scenario template includes a “Home Page” and “3 categories” section. If 

more categories are needed, it’s easy to simply add new categories into the excel 

spreadsheet. 

Go to AT Internet’ web site and download the Media Opportunity Calculator Excel 

Spreadsheet (complementary to this white paper). 

Media 

- 17 - 




About AT Internet 

AT Internet is a leading Independent web and mobile analytics solution provider, established 

in 1995 and with offi ces in 9 countries internationally. AT Internet prides itself 

on its customer centric approach and is able to boast 3,500 clients many of whom are 

widely recognized brands. 

Currently measuring 350,000 websites, AT Internet has several awards and distinctions 

to its name including the 2009 Platinum Distinction for European Seal of E-Excellence 

2009 awarded at CeBIT in Hannover, these distinctions underpin AT Internets core values 

of trust and integrity. 

AT Internet off ers a robust and reliable best of breed SaaS web and mobile analytics 

platform and partners with industry leaders to off er a complete solution to enhance 

your marketing intelligence and business eff ectiveness. 

Media 

- 18 - 



Headquarter 

le lafayette. 85, av. JF Kennedy 

33700 Mérignac 

France 

(+33) 55 79 23456 

France (Paris) 

20 rue Royale 

75008 Paris 

France 

China - Shanghai 

C/O Iptonic 

Room 102, NO.16, Lane 221 

Yanmin Road, Pu Dong District 

Shanghai China 200125 

(+86) 2 150 393 962 


Spain 

Avenida de Concha Espina 8, 1ª Izq. 

28036 Madrid 

España 

(+34) 911 105 829 

Germany 

Stiglmaierplatz/Dachauer Str. 37 

80335 München 

Deutschland 

(+49) 89/ 54 55 83 83 

China - Beijing 

C/O Iptonic 

Room 3005, Building C 

48 ZHongguancun South Street 

Hai Dian District, Beijing China 

(+86) 10 821 901 81 

Canada 

C.E.I.M. 

33 rue Prince, Montréal, 

H3C 2M7 - Canada 

(+1) 514 658 3571 

UK - Ireland 

85 Clerkenwell Road 

London EC1R 5AR 

United Kingdom 

(+44) 20 7681 4058 

Contact 

- 19 - 

White paper AT Internet 

AnalyzerNX is edited by AT Internet. Head offi ce: 85 avenue JF Kennedy 337000 Mérignac France - RCS Bordeaux B 403 261 258 - [DE.CL.8-000000583 - v1.0 - 08/2009]. Non binding documents and informations being able to evolve without advance 

notice

AT Internet

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

Delete template?

Save as template?