Facebook’s secret ‘Loyalty Prediction’ ad tool anticipates future user behavior & purchases

On the heels of Congress trying to determine how Facebook is managing user data, a new report from The Intercept claims the company is currently promoting a “Loyalty Prediction” ad service tool that lets advertisers target users based on predicted behaviors versus past actions.

Facebook’s predictive ad service tool will make it possible for advertisers to build campaigns beyond user demographics and preferences, instead using data that anticipates user behavior and purchases, according to a “confidential” document obtained by The Intercept.

“One slide in the document touts Facebook’s ability to ‘predict future behavior’ by allowing companies to target people on the basis of decisions they haven’t even made yet,” writes The Intercept reporter, Sam Biddle. “Facebook explains how it can comb through its entire user base of over two billion individuals and produce millions of people who are ‘at risk’ of jumping ship from one brand to a competitor.”

Knowing what we know about Facebook’s advertising model — and its ability to connect advertisers with relevant audiences — it’s a small leap for Facebook ad campaigns to go from targeting engaged users to targeting users who may be about to disengage with a brand.

The technology is reportedly connected to Facebook’s FBLearner Flow, an artificial intelligence-powered prediction engine first announced in 2016, although how it’s connected is unclear. The Intercept reports Facebook denied FBLearner Flow is used for marketing applications, but that it does use machine learning for ads.

In May of 2016, Facebook published a post on its engineering blog about FBLearner Flow and posted about the technology on its Facebook Engineering profile page, calling it Facebook’s AI backbone.

A Facebook spokesperson told The Intercept that Facebook uses “… FBLearner Flow to manage many different types of workflows” and that “machine learning is one type of workflow it can manage.” While Facebook denied FBLearner Flow is used for marketing applications, it’s difficult to separate it from machine learning that drives Facebook’s “Loyalty Prediction” tool.

The Intercept’s story notes the “confidential” document does not define the exact user information Facebook is indexing through its prediction engine for the “Loyalty Prediction” service, but it does list location, device information, WiFi network details, video usage, affinities and friendship details.

Facebook’s user data privacy issues have been around for as long as the site has had users. With the recent Cambridge Analytica firestorm, the company has once again been put under a microscope for not only its management of user data, but who it is opening its data up to — specifically, analytics companies now charged with using Facebook user data to manipulate elections around the world. A “Loyalty Prediction” tool raises the very same ethical issues, but on a much deeper, more complex level.

With the “Loyalty Prediction” tool, Facebook is not only curating thousands of data points across its user base to serve up ad audiences, its feeding those data points into a machine learning system that will anticipate what the next data point will be — a stark difference from simply collecting user data. If Facebook tells a company it can deliver relevant audiences, as well as predict which users are on the verge of not being relevant — it’s in Facebook’s best interest to do what it can for the advertiser and turn those users back toward the brand.

Considering that Facebook’s 2014 mood experiment on US users proved it could change people’s emotions based on the content it showed them, it’s not a stretch to see how Facebook could use what they know about altering user moods to deliver “Loyalty Prediction” results.

Marketing Land reached out to Facebook about its “Loyalty Prediction” ad service but has not received a comment.

About The Author

Amy Gesenhues

Amy Gesenhues is Third Door Media’s General Assignment Reporter, covering the latest news and updates for Marketing Land and Search Engine Land. From 2009 to 2012, she was an award-winning syndicated columnist for a number of daily newspapers from New York to Texas. With more than ten years of marketing management experience, she has contributed to a variety of traditional and online publications, including MarketingProfs.com, SoftwareCEO.com, and Sales and Marketing Management Magazine. Read more of Amy’s articles.