Gilad Barash, Brian Dalessandro, Claudia Perlich, Lauren Moores and Troy Raeder – all from marketing tech and targeting firm Dstillery – outlined this perspective in Cross-channel measurement and optimization, a paper forming part of the Advertising Research Foundation's Experiential Learning series of articles.
Using clickthroughs to establish the performance of mobile ads, they warn, is little more than a "fallacy". The idea that clickthroughs offer an insight into brand affinity is, equally, "not very likely" to hold true in practice.
In a hint as to why, data covering two weeks of mobile ad impressions for hundreds of brands showed the inventory concerned was a stronger factor at work than brand creative in terms of driving clicks.
"When taking a qualitative look at some of the inventory, we find that many of the most clicked on mobile apps are flashlights and games that appeal to children," the authors added.
"This begs the question on whether or not many mobile clicks are intentional, and whether or not the clickers are even adults."
Monitoring app and desktop usage, their analysis continued, is perhaps a more robust way of starting to segment the mobile audience in order to identify those individuals who are positively engaging with brands.
Dstillery premised this assertion on an experiment involving 25 marketers, and research that connected a week of mobile in-app requests from online ad exchanges with cookie-based brand engagements occurring on desktop computers.
Its "Brand App Index" determined whether the user of an app is more or less likely than the general population to take a "brand action" on desktop – like visiting a product's website or clicking on its content.
Having found "directional consistency" on this metric, meaning certain apps over-indexed in predicting this behaviour, and "rank consistency", so the results held true over time, the authors then proposed rating apps by category.
Prioritising a category of apps – like weather or navigation – in this way effectively becomes a form of behavioural targeting, and offers a stronger indicator of brand preference than a "proxy" like clickthrough rates, they concluded.
Data sourced from ARF