Whether your “friends” on Facebook are truly friends or simply “connections” depends on how you use the service. But if you’ve accepted a lot of “friend” requests, you know that the term really means “somebody connected to somebody else who somehow got connected to me.” That can be kind of fun because you’re exposed to new things. On the other hand, it also means you’re getting tons of posts about the activities of people you’ll probably never meet. Now Facebook is figuring out how to use that odd situation to its advantage through a new kind of search.
Graph Search is both a frontal assault on the other big search engines, specifically Google, and a way of unlocking value from Facebook’s 1 billion users, who have created what the company claims are a whopping 1 trillion connections to each other. We’re in the realm of “social search,” which means developing algorithms that let computers see connections between people and what they do. If Facebook gets this right, it opens up a lot of value for advertisers trying to target specific audiences, thus boosting its bottom line.
The advantages of ‘likes’
Run a typical Google search, and you’re running your question against a vast database of stored information, none of which has any necessary links to you. Social search would like to mine not just impersonal data but data generated by people you know, i.e., those you’ve connected with on Facebook. Beyond that immediate circle, social search opens out to the world of connections people have freely established between themselves on the service through places they’ve talked about, photos they’ve taken, and various interests they’ve linked themselves to.
In the early iteration of Graph Search you could, for example, search for “Books my friends read last year” or “nearby places to buy fresh coffee.” You could locate coffee roasters using Google, of course, but Graph Search would factor in the personal links so that you would be getting recommendations from your friends. It’s drawing directly from your Facebook identity so things you and your friends have “liked” by clicking the ubiquitous “like” button factor in. Need a plumber? You could search for “plumbers my friends like.”
The disadvantages of ‘likes’
This idea of personal recommendation has much to recommend it, of course, and it draws on the massive amounts of information people have fed Facebook’s database over the years. But it’s not necessarily an accurate source of information. Those “like” buttons, for example, are often used cannily by marketers so that to get a free service, you have to click the “like” button for a particular product. If you’ve ever clicked a “like” button to enter a sweepstakes or get your name into a contest, for example, it hardly counts as a real passion of yours being expressed.
So this is a mixed bag. I’m a jazz buff, and I’ve clicked to show my appreciation for various musicians over the years. That’s a sincere opinion and, in Graph Search terms, could be useful to someone else looking for something to listen to. But a listener who “liked” a band because doing so might have meant a free ticket was just being used by marketers, and the fact that I can now uncover these various “likes” through a quick search adds little value to my musical life.
Which makes the point that Graph Search is a work in progress. It’s still in beta, and you can sign up for early use at facebook.com/graphsearch. But the plan is to keep tweaking it, gradually adding all the posts that show up in your news feed and the data from your various apps.
Did Facebook worry you about privacy before? Well, a fully tuned up and distributed Graph Search is going to make you more visible than ever. So, are you thinking as carefully as you should about what you post on Facebook?
Paul A. Gilster is the author of several books on technology. Reach him at firstname.lastname@example.org.