Why Facebook’s Search Engine Won’t Be Anything Like Google’s
When Facebook founder Mark Zuckerberg mentioned during an interview last month that he wanted to build a search engine, headline writers instantly put Google on notice. Yet, while Larry and Sergey are probably watching closely, the technology and data at Facebook’s disposal suggest the company will most likely create something fundamentally different from Google’s search service.
Facebook lacks the comprehensive index of the Web that it would need to equal Google’s ability to match queries with Web pages—and it would have to invest a lot to create one.
However, flush with cash from its IPO this summer, the world’s largest social network already has its own unique stockpile of data—courtesy of its users’ social lives—that could power a new kind of search engine altogether. By mining users’ updates about vacations, music listening interests, online habits, and more, Facebook Search could be better at answering subjective questions, about what products, experiences, and businesses you might be interested in, than a traditional search engine.
“It would be very hard to create a general search engine to match Google,” says Apostolos Gerasoulis, a professor at Rutgers University who helped lead work on search technology at Ask Jeeves after the company acquired his search engine Teoma in 2001. Trying to replicate Google’s approach would require Facebook to spend considerable sums developing and deploying software “bots” capable of crawling billions of Web pages every day to gather a comprehensive index of the Web, he says. “Because Google is so big,” says Gerasoulis, “they have data for the long tail”—the uncommon queries for which relatively few pages are a match.
Microsoft’s experience with Bing should caution Facebook against such an approach. Since 2009, the Redmond company has spent more than $5 billion on Bing, according to some analyses. Although the quality of Bing’s results come close to Google’s by some measures, Microsoft has struggled to turn Web users’ heads. It serves only 15 percent of U.S. searches, compared with Google’s 65 percent.
A different approach may be more appealing to users of Facebook and other websites too. The social network has amassed a huge amount of data (see “What Facebook Knows”) because, in a sense, its users are crawlers that index tiny fragments of both the Web and the offline world. As well as recommending Web pages, videos, and songs by sharing them with friends, and labeling those recommendations with relevant descriptions, Facebook users check into restaurants and other businesses, and post photos tagged to real locations.
Gerasoulis says that could be the feedstock for a search engine focused on answering queries about the things that people share and discuss on Facebook, such as vacations, movies, recipes, and more. “When you go to specific subjects, the signals Facebook and other social networks have are amazing,” says Gerasoulis. That approach would also open up new avenues for advertising revenue, since Facebook could sell ads that appear next to the results for particular search queries. This is the very model that provides most of Google’s revenue.
Delivering on that potential would require sophisticated algorithms capable of weighting social information, says Gerasoulis. Google and Microsoft have both experimented with such things through their efforts to introduce social signals into their search engines (see “Social Search Without a Social Network” and “Why Bing Likes Facebook”). But Facebook has much more social data to work with than just counts of “Like” button clicks.
Mining users’ comments could help Facebook unlock even more useful data. The new social search engine Trove—built by a startup that just began publicly signing up users—hints at the potential of this approach. It can retrieve content scattered across a person’s multiple online accounts. For example, a search for “cute puppy” could reveal an unlabeled Instagram photo of a new pet because the photo previously elicited a tweet using the word “dog” and a Facebook comment saying “adorable!”
As the hosts of so much valuable information, “Facebook and Twitter both have teams working on search,” says Seth Blank, Trove’s founder and CEO. Digging deep into social data can uncover a wealth of information and forgotten content related to things people care about, he says, most of it not accessible by conventional search engines.
“If you’re planning a vacation somewhere, the truth is your networks have probably already discussed it at length,” says Blank by way of example; the networks he means consist of friends of friends as well as direct contacts. Blank believes his company will survive alongside a Facebook search engine by offering a neutral service capable of linking together different social sites. So far, the big social networks have been happy to let Trove work toward that, he says.
As Microsoft discovered, though, technology alone may not be enough to tempt people to try a new search engine. Facebook’s site already offers a search box at the top of every page, but people use it primarily to find other people, not search for content or answers to questions. Blank says research at Trove has shown that some people presented with a search box plugged into their social networks struggle to think of what to search for.
Gerasoulis says that is not an insignificant challenge for Facebook. “Search is about what you want right now,” says Gerasoulis. “You go to Facebook and hang out; it doesn’t currently have the same directness.”
If Facebook wants its search engine to succeed, it will need to craft something that not only is matched to the data the company holds but makes it clear to its millions of users why they need another search box in their life.
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
Deep learning pioneer Geoffrey Hinton has quit Google
Hinton will be speaking at EmTech Digital on Wednesday.
Video: Geoffrey Hinton talks about the “existential threat” of AI
Watch Hinton speak with Will Douglas Heaven, MIT Technology Review’s senior editor for AI, at EmTech Digital.
Doctors have performed brain surgery on a fetus in one of the first operations of its kind
A baby girl who developed a life-threatening brain condition was successfully treated before she was born—and is now a healthy seven-week-old.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.