Facebook Delves Deeper Into Search
On Feb. 1, a few hours after Facebook declared its intention to raise $5 billion in what will likely be the largest initial public offering in tech history, Mark Zuckerberg gave close followers of his company a potential clue to its future. On his Facebook profile, he uploaded a photograph of his desk and a large sign that read in big red letters, “Stay focused & keep shipping.” Yet it was the adjacent MacBook laptop in the image that drew the most attention. Visible on the computer’s screen was a blurry image of a Facebook page and, at the top, what seemed to be an unusually elongated white box. Web pundits speculated the image showed a prototype of a new Facebook search engine.
To date, Facebook hasn’t made search a priority, and it shows. The prominent white box at the top of each page is good at helping users find other members. It’ll also spit back Facebook pages for brands and locations, recent status updates from friends, and general Web search results powered by Microsoft’s Bing search engine. It’s a crude tool, however. Type in “Sonoma winery,” for example, and you get a disorganized assortment of wineries, people who work at wineries, unrelated banner ads, and a page for a wine-tasting iPhone app. In February, Facebook fielded 336 million search queries, according to ComScore—magnitudes fewer than Google and its closest competitors.
Searching the social network could get a lot better in the near future. About two dozen Facebook engineers, led by a former Google engineer named Lars Rasmussen, are working on an improved search engine, say two people familiar with the project who did not want to be named because the company is in a quiet period ahead of its IPO. The goal, they say, is to help users better sift through the volume of content that members create on the site, such as status updates, and the articles, videos, and other information across the Web that people “like” using Facebook’s omnipresent thumbs-up button.
The $15 billion search advertising market could be a huge opportunity for the company. It’s also a way to attack a chief rival, Google, which is moving in the opposite direction, from search to social, with its incipient Google+ network. With a more potent search engine, Facebook’s wine-loving users might be able to query the closest wineries that have been liked most often. That would give people one fewer reason to leave the site’s walled garden. Facebook could also follow the lead of companies such as Google and Microsoft and start selling relevant—and profitable—keyword ads alongside results. “Search is the best form of monetization on the Web by far, and they are leaving that on the table,” says Doug Leeds, chief executive officer of search engine Ask.com. “From a business perspective, you have to think about going into search.”
Facebook is unlikely to go toe-to-toe with Google for algorithmic supremacy. While Google controls 67 percent of the search market in the U.S. and has sophisticated technology to track a trillion Web pages, Facebook employs few, if any, traditional search engineers, who typically have deep expertise in fields such as information retrieval and natural language processing. But the company has a lot of social data it can apply to the problem of organizing information. Instead of crawling and ranking the whole Web, as Google does, Facebook already allows users to avidly flag the most interesting content, such as the best articles, recipes, and shopping deals. Improving Facebook search in some ways means just making more effective use of that data.
Google has recently attempted to integrate social data from Google+ into its results, but many critics consider it an awkward step that dilutes accuracy. Facebook may have more success, given its natural strengths in adding social elements to the Web. Gil Elbaz, CEO of data-crunching startup Factual and co-creator of the business that became Google’s highly profitable AdSense network, says Facebook’s data about its users and what they like could prove important. “Over time, this will let them build a powerful structured search engine,” he says.
Rasmussen, the Google veteran, is an interesting choice to help Facebook elevate its search game. A Danish computer scientist with a salt-and-pepper goatee, he co-founded mapping software company Where 2 Technologies, sold it to Google in 2004, and helped create Google Maps. Rasmussen later went on to build, with his brother Jens, the online communication and collaboration tool Google Wave, which was criticized for being too complex and was shuttered. Rasmussen jumped to Facebook in 2010 after a personal pitch from Zuckerberg. He told the Sydney Morning Herald that year, “I do think that social is a significantly less explored area still than search, and it is sort of the frontier of technology in many ways. But that doesn’t mean in any way that search is obsolete or even close to being obsolete.” Rasmussen’s brother remains at Google, adding a sibling rivalry twist to the competition between the tech giants.
A Facebook spokesman declined to comment on what Rasmussen is working on. If the engineer is looking to make the search box on Facebook more useful, he will face plenty of obstacles. Users see and rate only a fraction of the Web’s content—mostly material that is new—so Facebook may have a difficult time including more obscure content in search results. The site also continues to work with Bing and has to avoid ruffling feathers there. Zuckerberg meets every few months with Qi Lu, president of Microsoft’s online-services division, says a person familiar with the partnership, and in 2010 Bing began personalizing results based on what a user’s Facebook friends “like.”
Still, the search opportunity may prove irresistible. After Facebook goes public, likely in May, Zuckerberg will have to contend with investors who want him to cultivate new sources of revenue. Greg Sterling, a senior analyst at Opus Research, says Facebook could quickly become the second-most popular search engine if it tackles the problem in earnest. “There’s a huge amount of revenue waiting to be unlocked if they want to explore search-based pay-per-click advertising,” he says. “They can leverage the data and demographic information they already have.” Only then will that blurry photo of Facebook’s potential new search box come into focus.