Video search and recommendations startup Clicker announced Wednesday it has integrated with Facebook, giving it access to the social graph and helping it serve up personalized recommendations based on user interests and the interests of their friends. But really, how useful is the social graph in serving up personalized recommendations?
Using social data from Facebook, Clicker.com has redesigned its site and launched “Clicker Predict,” a video recommendation engine that helps it move beyond search to offer suggestions of movies and TV shows users may want to watch. When a Facebook user first logs in to the site, he or she now receives personalized recommendations based on what they’ve liked on Facebook or other sites around the web.
The deal is a major coup for Clicker, which can now use data from Facebook users’ behavior to offer a personalized “stream” of content users might be interested in. More importantly, it can now do so without users ever having to explicitly input data into its recommendation engine through ratings, comments or watching videos they’ve found on Clicker.com. By using the social network’s data, Clicker’s recommendation engine can classify or stereotype first-time users based on their Facebook interactions.
When asked how important the Facebook integration is in serving up recommendations in a phone interview, Clicker CEO Jim Lanzone said, “It’s especially important for early users, because now we have an idea of what you like from the start.”
The Facebook data is only one part of Clicker’s overall data set; the startup says its recommendation engine takes into account more than 50 different factors when it suggests videos, including user ratings and interactions with the site. In other words, Facebook data is just the starting point for personalized recommendations. As time goes on, and as Clicker collects more information about a user’s likes or dislikes, the Facebook data becomes less important, because it has a larger set of information to work from.
The question is how good social data is for making user recommendations at all. Netflix, (s NFLX) for instance, offered the ability to connect its subscribers with their friends to see what they were watching or what was in their queue, for year. But as time went on, Netflix made less use of this data and has since taken away the ability to connect with friends altogether. Subscribers to its new Canadian streaming service don’t even have an instant queue to guide them, but are reliant on the recommendation engine as the basis for surfacing new content they might be interested in.
Lanzone still believes having social data is important to serving a larger set of users. “For us to be able to read your mind, the more data we have, the better,” he said. “Our engineers would probably tell you that Netflix would be much better off having [social information], but what Netflix has is a much larger history with a larger set of users.”
Then again, Netflix might agree with him. Despite shutting out social connections between users in the past, the company is now hiring a Facebook engineer, according to a job posting. Though it’s not clear whether it, too, will look to use Facebook data for its recommendations.
Related content on GigaOM Pro: (subscription required)