Facebook released new plugins and raw data last Wednesday at its f8 developer conference, making it easier for developers to create apps based on larger amounts of data. The more ambitious experiments on this front will likely take longer to materialize, but here’s a quick look at what developers have already been able to do in the past few days since the launch.
Note that there are a few different ways developers are experimenting. The simplest way is to create aggregator and search sites for publicly shared data; the more interesting but difficult options means accessing Facebook Graph API data, then analyzing the data and combining it into other aggregator and search services.
Before we look more closely, here’s some context. Although Facebook has made some user data available to third parties for years, it has done so through Connect and its application platform. The Open Graph data makes already-public data like friend lists much more easily available. Its terms of service have changed to match. Since December, Facebook has been aggressively pushing users to share more data, making some criteria like names, profile pictures and friend lists public by default. Most recently, it required users to select Pages for interests, education and other aspects of their personal profiles — and because that Page information is public, that user data is now, too.
The cost has been some upset users, privacy groups and legislators (by the way, if you want to see what you’re already sharing publicly, check out this service).
However, Facebook has made more user data than ever available for other companies to build their businesses on top of, so we should begin to see more fruits. Facebook wants to make itself more important to other businesses around the web, better accomplishing its mission of being a social layer that both extracts value from and gives back to the web ecosystem. Of course, how Facebook controls access to its user data gives it leverage over companies that decide to use the data going forward. As with any platform, developers and entrepreneurs should carefully weigh the costs and benefits of using data from Facebook.
Aggregation
Two sites are aggregating shared items from Facebook’s Recommendations Plugin across online media web sites: Like Button and ItsTrending. Both show the same data in quite similar interfaces, displaying what your Facebook friends are “liking” and otherwise sharing on different sites, as well as overall numbers for how many Facebook users have shared each item. They are able to do this because Facebook lets developers access publicly-shared data by users and their friends, based on each users’ social graph; however, you’ll need to be logged in to Facebook in order to see personalized results on these widgets.
Aggregation in general has proven to be a decent business. Sites like Digg and Reddit use humans to do the aggregating — these plugins give a broader look at what’s being shared than those sites simply because Facebook is so much bigger. Yet those sites have active communities who are focused on finding and sharing news; although Facebook duplicates some of this functionality, it has so far been focused on private sharing.
More sophisticated aggregation could allow these new companies or others to better compete against the human-powered aggregators. But given the simplicity of the implementation, it’s not clear how any particular one will break out in appeal to users.
Another interesting example of an aggregation business is Tweetmeme, which looks at public tweets (which are most of them). Although its core service uses the same data that anyone else has access to, it has managed to break out and grow along with Twitter, and it has built out its service to be a main way for users to find top tweets by category. Perhaps Facebook plugin aggregators could provide similar refinements?
Beyond plugins, the Graph API data could be used to better tune what items rise to the top of their front pages. Combining active Digging with aggregated data from Facebook might help tune the top stories to a more general audience, for example. For an example of this, check out Fwix, a news site that has just incorporated some Facebook data to help people create and share customized categories of news.
Facebook has made some of this data publicly available since last fall, including public likes and shares. But that effort illustrates how its value depends on the developers’ needs. For example, news aggregator Techmeme, which algorithmically scours sources including content sites and other aggregators to figure out the top stories of the moment (with the help of human editors), has been using the data to help tune the top news that its sites feature . However, founder Gabe Rivera tells us that this Facebook data has had a “small impact” on how the company determines the value of stories.
Search
While aggregation is meant to bring the information that matters to users, search is intended to help them find what they want for themselves. Search is an obviously huge business, with technology market leader Google and smaller competitors like Yahoo and Microsoft trailing behind.
Facebook has been making data public intentionally so it will appear in search engines, as a way to play a bigger part in this market. The privacy changes have already increased the number of user status updates in search results, but so far Facebook has been providing the data manually to big search companies through business deals, including ones with Microsoft and Google. The Graph API now enables far more searching options, as we covered in more detail here.
Simple tools already take advantage of this. A prototype-style site called Open Facebook Search let you search anything public that Facebook is tracking from outside the site. You don’t need to be logged in to access the data.
A more interesting use of the Graph API is from OneRiot, a company that aggregates real-times searches. From the company blog, yesterday:
Until today, we’ve been indexing the links shared on Twitter, MySpace, Digg, Delicious and by our own OneRiot panel to help determine our search results. Now, with the addition of Facebook data, OneRiot delivers search results that reflect the pulse of a much, much wider social web.
The screen shot below shows what this looks like on OneRiot search results pages. You can see that I searched for “iPad” to find the latest buzz (this screen shot was taken last week). Because our results are informed by what links people are sharing on the social web, we attribute each result to the person who shared it first. We also indicate what service they shared it on. You can see in the screen shot that the third result was first shared publicly by a user on Facebook.
We expect search engines large and small to continue expanding their use of publicly-available data, as they have in the past year. Social data from Facebook could help improve the social relevancy of everyone’s results. Facebook itself could become a tighter part of the Facebook ecosystem. For more, be sure to check out our in-depth analysis of the topic in our Inside Facebook Gold service.
4.28.2010
Developers Experiment with New Search, Aggregation Products Based on Facebook Data
