Tech Industry

Yes, Google really should worry about Facebook

Google's algorithm connected people to information. Facebook's has traditionally connected people to people. But now Facebook is moving onto Google's turf.

The algorithm is the key to success.

That's how Google replaced Yahoo as the Web's best search engine in 1998. Google became the font of the online world's information by both finding more information online than any other search engine, and by figuring out what of it was the most important to the Web's users. Google algorithmically connected the Web to people.

Facebook, by contrast, has always been about connecting people to each other, but as the the latest version of the Facebook platform illustrates, the company is now about using that information to do what Google has traditionally done: connect people not just to each other, but to things, ideas, and media.

The algorithm is a big part of today's announcement at the F8 developers' conference. The algorithm can determine what you're likely to like based on who you like, what you do, where you go, which apps you use (and how), and so forth--all of which is information that Facebook will now collect through its own service and all the apps that are being built to run on it.

Mark Zuckerberg describes how Facebook will connect people to media based on the strengths of their connections to other people. James Martin/CNET

Google also knows what you do online, but it doesn't have close to the same depth of personal information that Facebook has, for two reasons. First, Google's core service, search, is a way station, not a destination. Google knows where you're going because you transit the site and because you may be tipping it off through your browser's search bar or through ads on the sites you visit. But Facebook is a destination. People go to Facebook and stay there. And communicate. And like. And so on. All the while, Facebook collects the data.

Second, Facebook knows who your friends are. In addition to the fact that you tell it this when you "friend" people, how and with whom you communicate on the site is more data that Facebook's algorithm can use to classify your connections to other people. When you respond to what Facebook is now calling "lightweight engagement" activities in the Ticker--when you decide to listen to a song alongside a friend, for example--Facebook files away this information, building, bit by bit, a dossier on your preferences and the people who are most likely to influence you.

For advertisers, this data is more valuable than Google's. Facebook will be able to cluster likely interest groups together and sell marketers access to those people. The company will be able to work with media companies to make advertising on their pages more effective. This is a serious and credible threat to Google's position as the Web's premier advertising provider.

For users, Facebook will, probably quickly, learn what each of us is likely to like by watching what we do on the site. This will help solve a big problem on a Web overloaded with novel information: discovery. By mining the "data exhaust" collected from the activities, links, likes, and so on that we all generate, Facebook should be able to predict, with increasing accuracy, what we're most likely to engage with, be it music or grocery ingredients.

Related reading
• What Facebook announced at F8 today
• Facebook unveils new look and feel (photos)
• Five things to know about the new Facebook
• Video: Zuckerberg shows Facebook's new look
• Yahoo News links with Facebook

If Facebook gets this wrong, users will continue to complain about the new design of the site as being too cluttered and confusing. But if the algorithm starts to feed people links to things they like but didn't know they'd like, it means the algorithm is working and Facebook is on its way to becoming the source of the most valuable information on the Web: who likes what, who they influence, and how to reach the people most likely to influence others (hint: go through their friends).

It's scary to see one single company own this database, but Facebook is coating this pill in sweet candy. We will find music we love through it. We will connect with friends to go on hikes with it. We will learn things from publications using Facebook because we see our friends reading them. And we'll make the whole thing easier for our friends, and Facebook itself. But stuffing Timelines full of personal resumes of preferences and activities.

There will likely be privacy missteps along the way, as Facebook turns on the algorithm and makes the data available to more developers through its platform. One might be tempted to step away from Facebook or to try hard to not engage with the flow of attractive links and media that comes through it. But I think it's going to be hard for people to say no to what Facebook will soon be offering.