Welcome to the Listening Post!

Online Audience Measurement: Beyond the Click

Tweet

 

Gian Fulgoni, Chairman of Comscore, presented “Measuring the Effectiveness of Online Advertising: Lessons Learned” at the 2010 American Marketing Association’s Marketing Research Conference, along with Todd Curtis, a senior consumer insights manager with General Mills. They discussed the decline of CTR (Click-Through Rates) as a measure of effectiveness, covering:

• The explosive growth in online advertising
• The challenge of accurately measuring its effectiveness; The C-level challenge – clicks and cookies
• One solution: continuously tracked online behavioral panels
• Lessons learned

Online advertising has 14.3% of the total, $22.7B in 2009, just behind newspapers ($24.6B) but far behind TV ($81.9B). The biggest piece of online advertising is search, at 47% of the pie, with display related advertising next at 35% (traditional banner ads, rich-media banner ads, videos and sponsorships). Display-related ad spending had been fairly stagnant from Q1 2008 to Q3 2009, before increasing in Q4 2009 to $2.3B with 20% year over year growth. Display advertising may be cannibalizing search, reaching people further down in the purchasing funnel than before.

Measuring advertising effectiveness is challenging. Click-through rates on individual display ads are “pitifully low” today, Gian said, averaging 0.10% (1 in a thousand people clicking on an ad). In the United States, only 16% of Internet users clicked on one display ad in the month of March 2009, down from 32% two years earlier. Yet click-through rates are the most used metric for measuring ad networks, used by 64% of 420 agency and advertising executives surveyed, with 61% using CPC (Cost Per Click). Interestingly, looking at it by the age of respondent, the older respondents are less interested in the click metrics. Only 35% say it is the “ability to drive brand awareness,” and 25% say it is the “ability to drive brand favorability.”

Gian said, “Clicks on display ads are a misleading metric and don’t reflect brand-building effects. Clicks don’t reflect the cumulative (latent) impact of ads.” Each month, 24% to 50% of Internet users delete their cookies, depending on geography. This makes it difficult to track a machine and overstates the unique visitors to a website by 2.5 times.

As an alternative to cookies, ComScore uses a panel of 2 million Internet users tracking their demographics, web visiting, online and offline transactions, search behavior, video consumption and ad views. ComScore panelists are divided into matched groups for those exposed to particular ad campaigns and those who weren’t. Even with low click-through rates, display advertising lifts web-site visitation, brand awareness and purchases.

Todd Curtis, with General Mills, provided the General Mills’ perspective as an advertiser. The digital ad development process should be the same as the TV ad development process, and the teams that follow a similar process have better results online. “We want to make sure our brands are where consumers are viewing media, and so we are increasing TV and digital advertising.” For some brands, the media plan is 100% digital, depending on the target audience, while on the other end of the continuum many brands are primarily focused on TV ads (though none have 0% digital). For online ads, which websites displayed the ads? How effectively did campaigns drive purchasing? “Who actually saw the ad on the computer? Was it a mom and her daughter or a man on the moon? We’ve had cases where we targeted women with kids and found that we hit men at 70% instead.” Some campaigns have been served outside the target country, at rates up to 33%, wasting ad dollars. Some campaigns have served up ads on adult web sites and other sites that conflict with the brand message.

As usage increases for different media, you need greater measurement precision. The key metrics for General Mills to measure online advertising are:

• Total impressions
• Unduplicated reach
• Frequency (5-10 impressions per campaign is ideal)
• Impressions by demographic (U.S. impressions)
• Test/control (creative is four times more impactful than other elements)
• Incremental consumer purchases

And General Mills is experimenting with others:

• GRPs: household, individual, targeted
• High value task
• Impressions by geography

More than 80% of online campaigns generate purchase lift, averaging 22% sales increase. Clearly, click-through rates are much less important than ever before when measuring online effectiveness.