How Facebook Graph Search will ignite a search revolution
Facebook's Graph Search is the future of search. Even before Google was a verb, the search engine Holy Grail was to deliver you the most relevant search results despite not knowing who you were and what exactly you were looking for. Now Facebook can stop guessing who you are—because it already knows you—and start serving up hyper-personalized answers tailored to you and based on the Facebook social universe.
Search leaders haven't been sitting idly by. Google's own hyper-personal search tool is called Google Now and landed on desktop search just last month. Microsoft's Bing has woven what it calls Social Search deep into its search engine. The hyper-personal search race has already been sparked; Facebook’s Graph Search ignites the revolution.
Personalized search is nothing new. We've gotten whiffs of the benefits of personalized search over time. Netflix has spent years honing its recommendation engine designed to keep you coming back to watch more movies and TV shows. Amazon recommends books, music, and numerous other products based on your past purchases. Pandora developed an algorithm that can generate playlists based on songs you tell it you like.
The secret to the success of Amazon, Netflix, and Pandora is the scope of the guessing was limited to you and a defined by a limited number of products, movies, and songs. The challenge for the leaders in search, Microsoft's Bing and Google, was that the data set was everything under the sun and you were an unknown. It's easier to create a search algorithm that uses past movie preferences to guess what similar movies you like. It’s much harder for Bing to guess what movie you'd like based on the query "find me a really funny movie that I'd like."
Now Bing, Google, and Facebook can begin to know who you are, who your friends are, your likes, where you go, about that failed diet, and where you vacation. The results are good, if we don't get too hung up on the privacy debate. In the age of Big Data, search engines can sift through your digital dossier and pair that with relevant search results.
But the bigger question is: How do (and will) hyper-personalized results differ between Bing, Facebook, and Google? Each service can't be the same by the nature of what they are, even if they all want to offer the same best search result.
Who knows best?
For example, ask Facebook's Graph Search for “friends of friends who have been to Yosemite National Park.” Theoretically, this query could link you to friends that might be able to give you tips on where to hike. You'd never be able to search Google and get a list of friend's names who have been to Yosemite, but you might be able to find better trails to hike that match your interests.
Facebook knows all about your personal relationships and your interests. Our early opinion of Facebook’s Graph Search proved underwhelming, but its potential is great.
Google knows about your Web habits, your most frequently emailed contacts, and your calendar appointments. A year ago it announced Search Plus Your World, a push to personalize your search results by including more Google+ profiles, business pages, posts, and Google+ and Picasa photos among the returns. Last month it began porting its Android 4.1 OS Google Now, a so-called intelligent personal assistant, to appear in desktop searches. Google Now is supposed to deliver useful information just as you need it, such as traffic and weather updates based on your search activities, location, Gmail inbox, and Google calendar appointments.
Bing relies on its partnerships to help personalize searches. And Bing's closest friend, thanks to a pricey investment, is Facebook. Bing harnesses the power of Facebook with its Bing social sidebar. Just this past week, Bing updated its sidebar with five times more Facebook data, Microsoft says. Bing social sidebar includes topically related status updates, shared links, and comments from Facebook friends. It also draws publicly shared data from high-profile users on other social networks, such as Twitter, Quora, Klout, Foursquare, and Google+.
Hitting the road
The wild card for personalized search is mobile. The biggest chance for hyper-personalization comes from the always-on and location-aware mobile devices we carry around with us every day. Market researchers at Comscore’s latest data suggest search is migrating away from desktops to mobile devices.
Facebook’s Graph Search on a phone may be years away. But one can imagine a mobile Graph Search app alerting us to, say, what percentage of Facebook friends liked a particular restaurant as we stroll past it.
Mobile technology both creates new data-collecting possibilities for search engines and allows them to be more situation-aware, delivering relevant results based on behavior patterns, the context of what you are doing, and when you're doing it. Google’s Google Now service says it “gets you just the right information at just the right time.” But in my experience with Google Now, it hasn’t yet delivered on that promise.
Try asking Google Now on your smartphone to “find me Starbucks” while navigating with Google Maps on a road trip. If you’re lucky, Google will find you a Starbucks just a few exits ahead. But it’s my experience Google Now more often than not chokes and spits up directions to a Starbucks I passed 20 minutes earlier.
I can’t decide who has collected the data about me. Is it Google or Facebook? When it comes to hyper-personalized search, maybe it doesn’t matter who has the biggest data set. Facebook’s beta version of Graph Search isn’t winning over the critics, yet. But the hyper-personalization search wars have just gotten started. By 2014, who knows: Maybe Facebook will find me a Starbucks 90 percent of my “friends” like, just down the road a few exits.
(PCWorld’s Tom Spring contributed to this report)