The failure that is Facebook Graph Search illustrates how much the company is sliding backwards rather than heading forwards. While the shortcomings of the new search process have been amply documented, suffice it to say that (a) it is still easier to find something on Facebook by entering the search parameter into Google along with the term “Facebook” and (b) an article on how to rejigger your privacy to meet the demands of Graph Search has been the most e-mailed piece on the New York Times website all week. (The fact that Graph Search surfaces the likes and dislikes of friends of friends-- along with their profile photos-- has a whole lot of people freaked out.)
Privacy is always going to be an issue for Facebook, but the main failure here seems to be in misunderstanding the effect previous “innovations” have on the accuracy of their information.
Take the ubiquitous “like” which replaced having to become a “fan” of a brand page. As this recent piece from Mashable pointed out, people would carefully curate the pages they were fans of so as show off their preferred version of themselves to the world: becoming a fan was a very conscious action. And while “like” was specifically designed to be an easier interaction that carried less weight, the result was that people used it promiscuously and with little forethought and/or curation, so that the list of things a user “likes” is rarely reflective of their actual tastes and preferences. Which leaves Facebook with data that’s fairly inaccurate and thus of limited value.
One might argue that Facebook Graph Search and the threat it offers to users privacy might be a spur for people to begin culling down their “likes.” That’s unlikely (no pun intended) as many users now have dozens, if not hundreds of “likes” and eliminating them will prove to be a rather onerous process. So Facebook is stuck with billions of halfhearted “likes” and no real way to do anything with them.
That’s almost, but not quite, as bad as the self-inflicted wound that is the Facebook brand page. For a while there, Facebook had a good thing going: brands had somewhat standardized pages with multiple tabs where they could run contests, share photos, handle customer service issues, etc. Even better, with a little work, those pages could be customized so that the graphics were consistent with the brand’s logos, not Facebooks.
Consumers liked the separation that brand pages offered, because, as I’ve been preaching since 2007, Your Brand Is Not My Friend™, something the current Timeline configuration ignores: with Timeline, brand messages are mixed in with friend messages and are promptly ignored as unwanted intruders. Compare that with the tabbed pages, where visiting a brand page was a conscious activity, something you did when you were in a mind frame to interact with that brand.
It’s the underlying idea behind the Ad Locker feature on the KIT Social Program Guide - brands benefit when people interact with them on their own time, when they are looking to do something that feels a lot more like shopping (which is fun) than listening to advertising (which is not.)
I get that Facebook is trying to find ways to monetize all the data they’ve collected from the third of the earth’s population who use the site. They just need to bear in mind that people don’t always act in a way that’s most expedient for advertisers. Build your model around the way people actually behave, and you’ll be able to find a workable solution.