Yosi Glick's Smarter Movie Recommendations

Movie recommendations suck. Or so says Yosi Glick, the founder and chief executive officer of Israeli startup Jinni. Watch a film like The Usual Suspects online, he says, and a lot of sites will recommend Se7en, another crime thriller starring Kevin Spacey. That kind of simplistic matching isn’t nearly specific enough, says Glick. The former is a twisting-and-turning heist story that audiences find clever. The latter is a disturbing, gore-filled tale about a psychopathic killer. Fans of the The Usual Suspects are far more likely to enjoy the Ben Affleck crime caper The Town than David Fincher’s dark detective story.

Glick knows this because Jinni has spent the last several years doing for film what Pandora did for music. The startup has created what it calls the Entertainment Genome, an online catalog of movies and TV shows described by thousands of parameters that determine whether viewers love or hate something. That includes obvious things such as cast and director, but also more nuanced data like the mood, look, and source material.

Glick left his job at Orca Interactive selling video-on-demand software six years ago to build up the team that became Jinni. A small group of film experts do some of the classifications manually, noting that the setting of the recent blockbuster The Avengers is “world-spanning,” for instance. Jinni’s experts have also developed software that analyzes online information, such as movie reviews and the data on entertainment sites, to tag flicks.

Users can test out the technology at Jinni’s website, but the startup makes its money by licensing software to big companies. Best Buy’s movie-rental website and Belgian cable provider Belgacom both use the Entertainment Genome to power recommendations. Microsoft (MSFT) struck a deal in September to use Jinni’s software in its Xbox 360 service, which offers movie rentals. Licensees pay Jinni a per-user fee. The privately held company won’t say if it’s profitable.

Jinni has also developed voice-recognition software that can interpret natural human speech. “You can speak to the TV to say, ‘I am looking for a funny romantic movie but not Woody Allen’s,’?” says Glick. LG Electronics, the world’s second-largest manufacturer of televisions, contributed to a $5 million investment round in Jinni in 2010 but won’t say if it plans to use the technology in future sets. “TV makers want to shore up their interface options before the arrival of Apple (AAPL) TV,” which is rumored to have voice-recognition capabilities and debut by 2013, says Gary Arlen, president of the media research firm Arlen Communications.

Glick says Jinni learns a user’s preferences over time. (Glick, for instance, is into “uplifting stories where the human spirit wins out.”) Eventually, the Entertainment Genome can be expanded to offer recommendations on books, games, apps, and more. “The goal is to build an entertainment personality for the user that spans all entertainment domains,” says Glick.

Earned a degree in information systems at Techion

Jinni's software detects the nuances of a film

Adding voice-recognition to its recommendations engine

View the original article here