Streaming Media - December 2007/January 2008 - (Page 54) the search is on Gotuit Takes a Different Metadata Route Unlike video search companies that look at existing metadata or have machines analyze the speech within the video, Gotuit relies on metadata generated by a human staff to help users find results across a specific, finite set of information, such as all the plays in the NFL for one Sunday or all episodes of Lost. The way Mark Pascarella, CEO at Gotuit, sees it, this approach enables users to pinpoint the exact information they want. “Today our company is delivering products with our partners across television, broadband and mobile and using the rich and granular level of metadata to enable an enhanced consumer experience where they can quickly and easily get to the parts inside the video they are most interested in seeing,” Pascarella explains. What’s more, this level of granularity enables advertisers to target ads more accurately, providing a way for Gotuit customers to monetize their investments. “By delineating each piece of video with metadata, we now have the ability to create highly targeted video advertising inventory and contextually the elements on the web page around the video player that can be targeted specifically to what’s being played back in the media player,” Pascarella says. Sprint is taking advantage of Gotuit technology to deliver NFL highlights each week aimed at fantasy football players using Sprint handsets. Pascarella points out that the NFL produces 40 hours of video each week, but most fantasy players simply track data and statistics. This technology allows them to see video clips associated with each player’s performance instantly without having to wade through hours of footage to find the desired 30-second highlight. To achieve this, Gotuit uses people to index the games as they happen, an approach that may sound time-consuming, but which Pascarella says the company has learned is much more accurate than machine encoding. For instance, he says a machine may not know the difference between Kyle Brady and Tom Brady (who happen to be teammates). To speed up the indexing process, they preload as much information as possible, such as teams and player names. Pascarella says they aren’t generating clips themselves, but instead they are marking different pieces with start and end tags and more specific tags, such as player names, to provide a way to traverse the larger data set. Therefore, if you want to view clips of Tom Brady, Pascarella explains, you enter his name and generate a list of his clips from the larger NFL highlights library. Pascarella says this technology can be applied to any type of video content, whether it’s the entire run of Seinfeld or last week’s National Hockey League highlights. Users can choose the path they want through the video library and customers can deliver targeted ads against those paths. Pull One Up. Talk with the people—more than 5,000 of them—who ARE the industry, and who use the Streaming Media Discussion Groups to network, discuss the latest industry developments, and get quick, knowledgeable answers to their business and technical questions. They’re free, and we don’t allow sales pitches. So what are you waiting for, an invitation? You just got one. StreamingMedia.com/Discussion 54 STREAMING MEDIA December 2007/January 2008 ESTABLISHED 1998 http://StreamingMedia.com/Discussion
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.