Natural Born Clickers Erin Hunter com Score Kim

Natural Born Clickers Erin Hunter, com. Score Kim Mc. Carthy, Starcom Greg Rogers, TACODA

Background

We knew something was happening when certain audiences continually performed the best on a Click-Through and Click. Conversion basis Click-Through Rates Click-Based Performance by TACODA Audience Click-based Conversion Rates Source: TACODA database of campaigns, October 2007

Sweepstakers skew heavily female and yet even for products that skew male, the sweepstakers will still click at the highest rates • • Sports cars Electric Shavers Golf Clubs Fishing Rods Sports Memorabilia Electronics Home Repair Male Pharma

Methodology

com. Score’s proprietary technology passively tracks all digital activity • Complete URL/clickstream data • No PII is retained • Very accurate knowledge of the user – Location – Demographics Panelist’s PC – Ad Exposure Browser – Click by Click HTTP Website CProxy Collection Server HTTPS Website Streaming Site

How com. Score tracks Natural Born Clickers • com. Score panelists browse online • All actual creatives load into the browser • Image URL, Server, Creative, Size, Type, Publisher • Includes cached ads • All exposures and clicks are added to the database • “Category” is based on “the page the ad was served from” Example: Clicking on an auto insurance ad for insweb. com on kbb. com (Kelley Blue Book) is put in the “Automotive Resources” subcategory *Adult content display ads were excluded from this analysis

How com. Score defined clickers • Clickers defined as anyone who clicked on at least one ad within the study time frame, July 2007 Clickers Non-Clickers Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

Online clickers follow the 80/20 rule • Divided clickers into the subgroups based on click volume: – Heavy: 4+ Clicks – Moderate: 2 -3 Clicks – Light: 1 Click • Despite only accounting for 6% of the total internet population, heavy clickers accounted for 50% of clicks in the month Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

Something to consider: • When optimizing your online buy to click based conversions, advertisers are letting somewhere between 6 -16% of the internet population dictate their online strategy

Who are Natural Born Clickers?

First let’s start with GENDER – there is little to no difference across sets of clickers Gender Indices against Total Internet Non-Clickers Light Clickers Heavy Clickers Index vs. Internet Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

But with AGE – the more people click, the more they will tend to be in the 25 -44 age range Age Level Indices against Total Internet Non-Clickers Light Clickers Heavy Clickers Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

And with INCOME – the more people click, the more they will tend to be in the lower income brackets Income Level Indices against Total Internet Non-Clickers Light Clickers Heavy Clickers Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

But yet, when we look at BUYING POWER INDEX, heavy clickers spend about 2 X what non-clickers do Buying Power Index Note: A BPI value over 100 means that a clicker purchases, on average, more (in dollar terms) than the average Internet user during the reported time period. Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

The counter intuitive spending pattern may be a result of the extraordinary consumption of the web demonstrated by the Heavy Clickers Total Minutes per Visitor Total Pages per Visitor 5 X Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period 8 X

Click-through rates show increased likelihood to click (2 x), even after controlling for consumption Click-through Rates Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

When we look at COMPOSITION INDICES by website category, we begin to see an internet of two drastically different lives Non-Clickers: Ø 68% of the Online Population Ø 0% of all clicks ØLike Portals, Search, News, Finance Non Clickers Heavy Clickers: Ø 6% of the Online Population Ø 50% of all clicks ØLike Gambling, Job Searching, Games Heavy Clickers Note: Indices are against each segment’s average UV composition over all categories. For example, compared to their average category consumption, Heavy Clickers are 40% more likely to visit Gambling sites Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

Something to consider: • Clickers in general are a unique group of people • Consume more online media than non-clickers • Demographics differ from general population • Entire web experience appears to be leisure time • Increased likelihood to click • As marketers we have become enamored with the behavior of Heavy Clickers – but are there potential consequences? • Can we find an environment where typically “normal” clickers enter into the heavy clicker mindset? • Find a category where a user set is spending disproportionate amounts of time • Look at how clicking behavior is effected

Social Networks

Why Social Networking? • Inherently leisurely category for teenagers – P 2 -17 over index (146) for total minutes spent on Social Networking sites • Overall category demographics differ from that of heavy clickers – Age skews younger – HHI skews higher – Geography skews are negligible • Despite the category differing demographically overall from the heavy clicker profile, a lot of clicking activity is happening here – Heavy clickers make up 9% of Social Networking Visitors (155 Index) – Social Networking Heavy Clickers make up 23% of total Heavy Clickers • Social Network Heavy Clickers defined as those who clicked on an ad within Social Networking content 4+ times in the month • Will heavy clickers look the same in this environment as they do overall? – Are Heavy Clickers the same in all environments? – Or will their demographics more closely mirror those of the site category?

Social Networking Heavy Clickers are almost twice are likely as Heavy clickers in general to be 12 -17 years old Age Level Indices against Total Internet Heavy Clickers Social Networking Heavy Clickers Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

However the income skew is consistent between Social Networking Heavy Clickers and Heavy Clickers overall Income Level Indices against Total Internet Heavy Clickers Social Networking Heavy Clickers Source: com. Score, Inc. custom analysis, Total US Online Population, XPC Persons Panel, July 2007 data period

Something to consider: • When we find a consumer segment in a leisurely mindset, clicking does increase – P 12 -17 clicking at a much higher rate than groups spending less time on Social Networking sites • However, income skew stays consistent, despite shifting age skew – Will this hold in other categories? – Is income a key driver in clicking likelihood?

Preliminary data suggests no correlation between clicks and brand metrics • Starcom Clutter and Context study examined the relationship between behavioral and attitudinal metrics – Site level data from Dynamic Logic Market. Norms and DART was matched – Sites’ average delta on each of five brand metrics was correlated to average click thru rate • As the below chart shows, no attitudinal metrics were correlated to click through rate • When digital campaigns have a branding objective, optimizing to clicks is not necessarily improving campaign performance

Conclusions • Assumption of click based optimization is that everyone that sees your ad is equally likely to click it, but data shows this assumption to be false – If anyone, we would expect those within your target to be the ones more likely to click – Specific group of people click on ads, often not within your target • When you optimize to click-based conversions you are shifting your plan to reach a specific type of internet user – “Casino” mentality • • • Spending a lot of time online Spending more money online Exploring more online – Unless you are targeting this type of person specifically, you cannot rely on CTR alone • Even when click-based conversion is your metric of interest, you are still limiting the reach of your campaign by optimizing based on the behavior of a small group – Return rate? Life time value? • More time = higher likelihood to click – When targeting busy demographics (e. g. c-level executives, moms), would expect lower click thru rates