Facebook may be entering the search race, but it's still far from overtaking Google.
The blogosphere is abuzz today over news that Facebook will start including some sites that use its Open Graph protocol in search results. Many have described the update starkly: Facebook "guns for Google," "brings the fight to Google," and, my favorite, "just threw down the Google-flavored gauntlet." It's significant news, but Facebook's approach suffers from the same features that make it so unique.
In April, Facebook introduced its Open Graph protocol, which offered webmasters two things: a "like" button to install on their webpage for Facebook users to click on to show their approval; and a set of tags webmasters can use to tell Facebook what their site is about. Now, the social network confirmed to online publication All Facebook that "all Open Graph-enabled web pages will show up in search when a user likes them."
Facebook can take a search for "Michael Jackson," for example, and return pages self-identified as being about the pop star and arrange them in order of most liked. Unlike Google and other search engines, which rely on links and other metrics to rank results, Facebook can outsource the identifying and organizing of sites to webmasters and its own users.
But Google can take advantage of those tags that webmasters use to tell Facebook about their sites because it's all in the source code of the page itself. So, that means that the only real edge Facebook has -- and it is a significant one -- is its database of "likes." Still, Facebook's new search suffers from some problems and shouldn't pose too much of a threat to Google.
First, only pages that use the Open Graph protocol will show up in a Facebook search, according to the All Facebook report. Anyone without a "like" button is excluded. That might add an incentive to add the buttons, but it also means Facebook is leaving out large swaths of the Internet. Also, companies and governments are often required to make information available that they would otherwise bury. They may lag behind on or never use Open Graph, leaving their content out of Facebook's reach. Facebook's approach is passive, the company only gets what sites want it to get, while Google's is active, it gets whatever it can find.
Second, Google also has its own (admittedly less accurate) way of gauging user interest: tracking clicks. Granted, everyone clicks on links that end up being useless, but the practice of tracking them still gives the search engine some rough sense of subjective interest. En masse, the useless ones can likely be weeded out over time when analyzed alongside other metrics such as how often those sites are linked to.
Finally there's also the fact that subjective interest just isn't always the best indicator of a good result. If I conduct a search for "BP oil spill," chances are the most "liked" links won't be the latest news stories that I'm looking for, but anti-BP screeds. The emotional response to any issue or individual could skew the search results away from the topic at hand and to the response.
Of course, none of this applies if Facebook decides that to add other metrics like Google's PageRank to its search results. If that happens, Google should be concerned.