New APPS

By Gordon Hull

One of the things that marketers like about big data is that they can personalize ads.  That operation is getting increasingly sophisticated.  We’ve known for a while that basic personality traits (like introversion/extraversion) can be predicted from Facebook likes.  I missed this paper when it came out, but some of the same authors as the initial Facebook “likes” paper have now done the inevitable follow-up.  By targeting Facebook ads on the basis of openness and extraversion (where the correlation with likes is fairly robust), they were able to make users 1.54 times more likely to make a purchase (thus about 40% more likely) than with non-“psychologically-tailored” advertising.  The study size, as with most research on FB, is enormous – the study reached some 3.5 million users (a point the authors use to try to defuse some objections).  The authors duly note the murky ethical issues that emerge: on the one hand, you could proactively try to assist people who show signs of depression; on the other hand, weaknesses like susceptibility to addictive gambling or dubious political targeting could be exploited.  That observation is pretty obvious now, in the wake of Cambridge Analytics.  Two other takeaways stood out more to me.  First, the study predicted personality on the basis of only one like.  As the authors note, that means the study likely underestimates the potential effect of psychological targeting.  Second, the authors emphasize that a point that a lot of us have been trying to make for a while: that these results undermine a lot of the current regulatory strategy for privacy. I’ll let them speak:

“The psychological targeting procedure described in this manuscript challenges the extent to which existing and proposed legislation can protect individual privacy in the digital age. While previous research shows that having direct access to an individual’s digital footprint makes it possible to accurately predict intimate traits, the current study demonstrates that such inferences can be made even without having direct access to individuals’ data. Although we used indirect group-level targeting in a way that was anonymous at the individual level and thus preserved—rather than invaded—participants’ privacy, the same approach could also be used to reveal individuals’ intimate traits without their awareness. For example, a company could advertise a link to a product or a questionnaire on Facebook, targeting people who follow a Facebook Like that is highly predictive of introversion. Simply following such a link reveals the trait to the advertiser, without the individuals being aware that they have exposed this information. To date, legislative approaches in the US and Europe have focused on increasing the transparency of how information is gathered and ensuring that consumers have a mechanism to “opt out” of tracking. Crucially, none of the measures currently in place or in discussion address the techniques described in this paper: Our empirical experiments were performed without collecting any individual-level information whatsoever on our subjects yet revealed personal information that many would consider deeply private. Consequently, current approaches are ill equipped to address the potential abuse of online information in the context of psychological targeting” (internal citations omitted).

Marketing has of course become increasingly targeted over time.  Early magazine ads imagined their readership into very rough categories (“homeowning housewife”) and tailored ads to an imagined construction of what such a reader was like; this later moved into empirical work on the tastes of such individuals at increasing levels of granularity.  But data represents something different insofar as it vastly increases the amount that can be known about targets, and it enables the sort of predictive modeling that the work on FB likes demonstrates.  That result was anticipated in a 2015 paper that proposed that systems could dynamically adapt nudges to the type of ‘persuasion principle’ that was most effective on the individual being nudged. Thus, some people would see a message instructing them what to do; others will receive one claiming that a social consensus supports the desired behavior, etc.   As Adam Arvidsson was able to presciently propose fifteen years ago:

“As consumers, mobile and supposedly self-reliant, we are subject to a virtually ever-present ‘panoptic sort’; a surveillant gaze that expands far beyond the walls of the prisons, factories and hospitals and schools where, according to Foucault the panoptic model first developed. The physical, social and cultural mobility of social life, the moving about between environments and activities that has become a key characteristic of post-modern life, has also become a source of value to be realised on the market for commodified information.” (457; internal citations omitted)

The current research on FB likes as a tool for targeting shows this is not that hard at all. You can find out everything you need to know about somebody based on what they say they like on Facebook.  One of the likes associated with openness?  “Philosophy.”  Enjoy your ads!