NEW YORK (AdAge.com) -- Launching its universal "like" button, Facebook extended its tentacles across the internet today, setting up pipes to gather user data from anywhere on the web. And now that users can add what topics, products or content they like to their Facebook profiles, the social-networking site is sitting on a data treasure chest.

At the F8 developers conference today, CEO Mark Zuckerberg announced a platform that aims to connect the entire internet through the social network. With those like buttons appearing on major publisher sites directly after the announcement, users can thumbs-up individual pages with one click and publish that to Facebook. Meanwhile, that Like is stored for later.

"[Zuckerberg] is using the like button as the glue to link Facebook to everything else and understand his users much more," said Shiv Singh, Razorfish's global social media lead. "It's a data goldmine."

Facebook's new tools, including the like button, activity feeds for other Facebook users and recommendation engines, are designed to embed Facebook functionality on outside websites. With like buttons on 75 sites, including publishers such as CNN and the New York Times, from day one, Facebook expects to serve more than 1 billion buttons in the first 24 hours. Once a user likes a page, the publisher gets a link on the user's page, and means to later publish to that user's newsfeed.

Facebook made no specific ad announcement today, but the affinity data for the site's more than 400 million users already has agency types salivating. A Facebook spokeswoman said its policy about developers or publishers targeting ads on their own sites has not changed with the new policies. Facebook will allow developers to apply user data to target ads on their own sites, but not elsewhere. Even though sites can't share data, Facebook will be sitting on the mother lode.

New giant
"Facebook potentially could power an all-knowing behavioral-targeting platform the likes of which we've never seen before," said Ian Schafer, CEO of Deep Focus.

Short term, the platform could mean more time spent on publisher sites and traffic from Facebook. Mr. Zuckerberg also hailed the "open social graph" as a way to create personalized web experiences where publishers will be able to tailor content to a user based on his or her like history.

With one launch partner, Pandora, a new user to the music site can automatically be served a playlist of bands he or she has liked elsewhere on the web. Then, once that user visits a concert site, band preferences are ported over from Pandora to tell that user when favorite bands will be playing nearby. Users who don't want this level of personalization can adjust privacy settings within their accounts.

But how expansive is access to that data? While it'll feed recommendation engines to serve users relevant content, its implications on ad targeting even within publisher sites are still unclear.

According to a blog post from Bret Taylor, co-founder of social-sharing startup FriendFeed that Facebook acquired last year, publishers will also get analytics from Facebook's insights product. Though one launch site, USA Networks, which started with embedding the like button, the recommendation engine and activity feed, is not yet looking at Facebook for ad targeting. Its main focus for now is turning fans into ambassadors for its shows.

Competing with Google
"This is about is getting our brands and our shows in the communication flow," said Jesse Redniss, VP-digital for USA Networks. "It doesn't have to drive back to our core property; we want to push our brand out." So far, access to Facebook's trove of data will include top-line psychographic information that will help serve relevant content to users. Beyond that, implications for ad targeting are not clear. "For us, we are not going to start data mining that information," he added. "We are more interested in user habits today."

Facebook also launched a documents product with Microsoft: Docs.com. It's the web version of Microsoft's Office suite, designed to share and collaborate online. The site appears to be coming head-to-head with Google Docs.

"Google just found its nemesis," added Mr. Schafer. "Instead of targeting people based on their click behavior or search behavior, it's targeting based on their relationship to people and to brands and content." And with relationship targeting instead of just contextual targeting, Mr. Schafer anticipates higher ad rates for publishers.

"They're both after the same behavioral information," he said. "Google has tried to capture information on social graph through Orkut [Google's social network] and Buzz [its Twitter-like service], but they just can't do it the way that Facebook does."

Targeting against likes will presumably be available from the get-go on Facebook.com, since the site already lets advertisers serve ads to users with specific interests or profile settings. Plus, with likes happening in real time vs. user profiles that often are years out of date, advertisers will get more current data. User endorsement coming in from outside the Facebook ecosystem will also prove valuable for brands and publishers.