Almost two years ago The Filter, a startup backed by Peter Gabriel, launched to bring better music and movie recommendations to consumers. The site got lost in the abundance of more popular music and movie sites out there, so about a year ago CEO David Maher Roberts decided to shift gears and start licensing his recommendation engine to other businesses.
It was the right move. Today, the Filter powers recommendations for sites and devices with a combined reach of about 20 million people, with two more large media deals in the final stages of converting from a trail to a full license which will bring its total reach up to 85 million. The startup’s revenues went from $150,000 in 2008 to about $1 million in 2009. “All that money came from licensing,” says Roberts. “I think we get $2,000 from Google for advertising.” Since November, the company has been “borderline breakeven.” And it just added to its board of directors former Google engineering VP Doug Merrill, who left Google to briefly serve as president of EMI for a year.
“Recommendations—from friends, from newspapers, from colleagues—are the most common way to find new content,” says Merrill. “However, there is more information available than there are people to recommend. The Filter analyzes data to provide measurably better, more relevant recommendations, automatically.” The Filter creates personalized music and movie feeds based on user’s activities (rating. listening, saving, sharing) as well as their preferences in other accounts such as iTunes and Last.fm. Roberts claims that in trials, customers have seen a 20 to 40 percent lift in media consumption (video views, dwell time) than using their own recommendation algorithms. Not counting those two big media deals in the wings, the Filter’s recommendation technology is currently being used by Nokia Music, Sony’s MyPlay, DVDPost (a European Netflix), ThePlatform, and We7 (another music site also backed by Gabriel).
The Filter’s own site is still growing steadily, if slowly, with about 800,000 unique visitors per month. “It was a much slower process than we had wished,” says Maher. People liked the technology, he says, but they wanted it on sites where they already consume content. The Filter’s experience shows how tough it is to build a standalone music or movie property. But if its recommendation algorithms really do provide the kind of improvements Roberts claims, more sites will adopt it. You can try out its relevance engine for yourself on The Filter website.