Will Mobile Move to a CPA Model?

I presented yesterday at a conference called SEMpdx, which is a search conference in Portland. My presentation was on "Mobile Marketing Strategies & Tactics." Reid Spice of iCrossing was my co-presenter. Prior to the session we had an informal discussion about mobile search, CPC pricing and "inadvertent clicks" on mobile ads. The latter was tied to a survey that got fairly wide (and uncritical) attention at the end of last month. The survey was sponsored by mobile lead-generation company Pontiflex and conducted by Harris (in December, 2010).

There were a wide range of findings from the survey, which was conducted either twice or in two parts and had more than 4,000 total responses from US adults. The survey asked a number of questions about apps, advertising and mobile user behavior. However the big headline and finding that got all the attention was "47 percent of mobile app users say they click/tap on mobile ads more often by mistake than they do on purpose."

Other survey findings included the following: 

  • Almost two-thirds of mobile app users selected ads that contain coupons, deals or newsletters as their preferred in-app mobile ad type
  • Mobile apps are popular across all age groups; usage is highest among users 18-34 years old

The press release pulled out a piece of segmented data that is even more pronounced: "61 percent of mobile apps users ages 18-34 click/tap on mobile ads by accident more often than on purpose." The "message" conveyed by these findings is that people frequently click ads by mistake and so clicks are a false metric and unreliable as a measure of the efficacy of mobile advertising. Let me say that I absolutely agree with that statement; clicks are a false metric that mobile marketers should walk away from. 

However there are a couple of problems here. Self-reported data and "behavioral data" are often inconsistent. In other words, what people say they do and what they actually do may not always line up. So as a fundamental matter we don't know whether people are in fact clicking by mistake more often than not. And we don't have a "control" or baseline metric for the PC to compare it against. It could well be that there's a high incidence of mistaken clicks online.

As an aside the audiences that click display ads online are "a small, unrepresentative portion of the total online population." According to comScore (2009), "4% of Internet users account for 67% of all display ad clicks." Thus clicks are a really bad metric for the efficacy of display ads on the PC. Yet lazy marketers still use CTR as a measure of ad effectiveness. 

Most of the coverage of the Pontiflex survey turned into a kind of "telephone" game that ultimately distorted what the publicly released data said. The coverage, and especially the secondary coverage based on the first stories, turned into: "half of all mobile clicks are inadvertent." 

In fact we don't have any sense of what percentage of mobile clicks, according to this survey, may be "bad." We know that 47% of these survey respondents think (key word) that more of their clicks than not are unintentional. It could be. But I'm surprised that nobody who covered this was critical or sought to drill into the data for more information. 

I don't disagree with the argument that Pontiflex makes about engagement, lead capture and other KPIs being better vs. clicks. And if these Pontiflex data are supported by other findings from others we could see marketers make a push toward CPA (publishers prefer CPM of course). Networks like OfferMobi are also pushing CPA as a better mobile metric too. One might argue that mobile is better suited to CPA than online in many respects because mobile user behavior is more "action-oriented" than PC user behavior. 

If CPA does become the standard in mobile display -- we'll probably see a mix of models however -- that, in turn, might influence online advertisers as well to demand or gravitate toward CPA. Search marketing, whether online or in mobile, will largely remain CTR-based however.