Wednesday, January 16, 2013

Facebook Graph Search: Necessary, But is it Sufficent?

Facebook seems to have done a good job wooing the press for the introduction of its new Graph Search feature, with dozens of articles strewn across the Web touting Facebook's new search feature as an awesome breakthrough in social media discovery. But let's be clear about what this feature really is, what it really offers and what it doesn't offer. Put these three things together, and I think that what you get is the story of a company that has lagged woefully in content discovery tools that is just beginning to catch up.

Graph Search is not really a revolutionary tool at its heart: it's a combination of relational database lookup and a natural language processor on its front end that organizes the database lookups. The fields that are searched are the relatively sparse personal information that a person or entity enters into their Facebook profile, plus the links and signals (such as "likes") that people enter into Facebook. This information is used to retrieve profile cards for the people and entities that match the query.

If you're saying to yourself, "Um, I think that we've seen these things before," well, you're right. It's been over a decade since search engines like pioneered natural language query on the Web, and social media sites such as LinkedIn have had far more extensive social network queries based on their more detailed profiles since its inception for about as long as that. So please, let's not get overwhelmed by information science that's old hat, for the most part. Moreover, the queries that you're targeting are based on people who are a part of your social network. So claiming that this exposes billions of people to a search is only valid if you have billions of people in your social network or who are interested in a particular topic. For the most part it's more akin to a sophisticated contacts lookup feature.

Moreover, it intentionally doesn't include content from the Web. As The New York Times put it in their review: "Its search tool is based on the premise that the data within Facebook is enough and that its users will have little reason to venture outside its blue walled garden. What they cannot find inside the garden, its search partner, Bing, a Microsoft product, will help them find on the Web." In other words, this is meant quite intentionally to infer that information about your social graph can only find relevance within the walls of Facebook's little world.

However, at least for the moment Facebook's feature does give it somewhat of a leg up in its growing competition with Google's growing Google+ community. Google's main search engine tops the Web, of course, and Google+ content from social contacts in Google+ are integrated into Google's search results by default. And this, too, has a natural language interface, one that isn't hobbled by trying to fake a formatted query via natural language. It's also worth noting that Google+ has similar lookups of one's social network with simpler queries. If I type in "music" into to the Google+ search box, for example, the autocomplete search feature automatically brings up a listing of post, people and communities in my network that relate to this topic. Add on a little natural language fluff and it could look just like Facebook's tool.

Upgrading Google+ search for even more powerful results wouldn't be too hard for Google to do, given that Google+ profiles have more in-service data fields than Facebook and it also has the benefit of pointing to more content on the open Web than Facebook. Unlike Facebook, Google seems to use Google+ more as a Web content relevance indicator than as a walled garden, so while its search is powerful, Google is not so intent on keeping people inside Google+ as a walled garden - they want you to find good content on the Web, because, after all, that's where their ads are and where their indexing strengths lie.

So there's no doubt that Facebook's Graph Search is a necessary upgrade, and a good one. But I have strong doubts as to whether it's going to be sufficient in and of itself to make Facebook that much more a valuable site. The data that it can leverage is powerful, but its inability to integrate effectively with people's Web presences and interests would seem to be a strong minus, one that's not likely to be addressed any time soon. Facebook is indeed trying to be a destination that obviates the need for the Web as much as possible, and in this it is just another "big media" play and not very social. Real social media acknowledges that we post things all over the place, and that this is OK. So kudos to Facebook for finally putting an indexed search on the Web site. Please let us know when there's a real breakthrough, though.
Post a Comment