There’s been plenty of buzz recently about some Facebook ad nonsense. Various sources are claiming that a new comScore study “proves Facebook ads work.” There are multiple problems with this assertion: 1) the study is not even about Facebook ads (at least as they pertain to broader demographic, rather than fan, targeting), and 2) like other studies, this one does not necessarily prove that Facebook engagement causes increased sales among fans, only that it is correlated with the trend. I’ve talked about the second point before, but it bears repeating: even if your fans buy more, that doesn’t necessarily (or even probably) prove that Facebook made them do it. Although the comScore tests made use of control groups consisting of non-fans with similar purchase patterns prior to Facebook engagement, future tests might to better to focus on a/b social promotions for specific goods among fans–a specific promotion’s ability to sell a specific product among fans (vs among fans unexposed to that promotion) would more directly prove the value of (unpaid) Facebook promotion itself, rather than the somewhat self-fulfilling prophecy of fans being more likely to buy in general.
On the heels of the widespread sloppy reporting on that study, though, is more positive news about Facebook’s app (and game) curation. Rather than emphasize the apps and games that are most popular overall, as many app stores do, Facebook will instead emphasize those that are most popular with your friends, and particularly with the friends you engage with most. As noted earlier, this may give rise to a filter bubble effect that precludes the discovery of apps outside of your typical interest base, but it’s a good step toward learning why certain apps appeal to certain people.
Unlike ads, apps provide true self-fulfillment: if you use the app, it has clearly been a successful, causative transaction, contained within the Facebook platform. Facebook’s timeline and app expansion are paving the way for a more direct relationship between Facebook engagement and purchases, increasing the ability to attribute causation more directly to the Facebook-purchase relationship. For example, Starbucks could create a Facebook app that lets users mix their own Frappuccinos–and then even purchase those exact mixes on Facebook for pickup in-store (or delivery!). Purchases of a specific product deriving from usage of a specific app would suggest a more direct causal relationship than simply being fans of a brand. And Facebook pages themselves are becoming more like storefronts as brands add apps related specifically to their value adds.
How would you design a study to directly prove (once and for all!) Facebook causation for purchases? Can it even be done? I look forward to haring your ideas–after I come back from getting an iced coffee!