Facebook Adds Algorithm, Indexing to Grow Search Capabilities

Facebook has always had its eye on search, but the company is trying again to take on Google and Twitter. Facebook already handles 1.5 billion searches a day on its site, but the vast majority of those searches are for names. In October 2015, Facebook made it possible to search for all public posts, developing an algorithm to rank trillions of posts daily. If its efforts pay off, Facebook will dwarf Twitter’s 320 million users, and provide a more personal, friend-oriented spin on Google’s more generic search.

Bloomberg reports that Facebook is betting on the idea that “the average Facebook post is interesting for longer than the customary five hours it might be immediately findable in the news feed,” be it chatter about political candidates, dinner recipes or vacations.


“Our founding belief is that there’s a huge amount of wisdom and knowledge in that information after 12 hours, after 24 hours,” said Facebook vice president of search Mike Vernal.

For it to work, Facebook needs indexing: every post must be scanned for keywords and such “qualitative” factors as writing quality and whether the author gets many likes and is an authority on the topic.

“In order to have Facebook have the best most accurate results, we need to make sure that all the updates — every single photo, every single video, link, share, like, comment — is reflected in our index in a matter of seconds,” explained Facebook director of engineering for search Girish Kumar.

Once people realize how their posts might be indexed, some might tighten up their privacy settings, but others might “scramble toward the limelight, crafting timely posts for a mass audience and giving the social network into the kind of topical and pulse-quickening timeline currently featured by Twitter.” Kumar notes that the algorithm prioritizes “results from original sources and authorities on the news” as well as friends, brands, pages, celebrities and strangers.

Already, anytime a user shares an article, Facebook uses its search algorithm to show how many people are talking about the topic and suggests related stories on the site. If a user “checks in” to any location via Facebook, the company’s search algorithm proffers posts from other people who have been there.

“If search on Facebook really takes off, I think Google is under pressure,” said Axiom Capital Management analyst Victor Anthony. “If they get it right and they’re able to monetize against searches, it’s extremely lucrative for Facebook — billions in revenue.”