Clerk Dogs: Recommendations By Video Clerks

By Geoff Duncan December 10, 2008

Clerk Dogs: Recommendations By Video Clerks

A couple years ago, movie rental service Netflix offered up $1 million prize to a software developer who could come up with a better way to recommend new movies to its millions of subscribers. Netflix wasn’t asking for much&mdah;just a 10 percent improvement—but to date the prize has gone unclaimed, despite more than 27,000 entries. (Although Netflix did give $50,000 to a team of researchers at AT&T who came up with an 8.5 percent improvement kast year—AT&T donated the money to charity.)

Now a new movie recommendation service called clerk dogs is taking the opposite approach: instead of basing recommendations on aggregate user data and viewing histories, clerk dogs taps into the collective wisdom of movie-watching professionals—a.k.a. video store clerks—to fuel its recommendation service. Founded by serial entrepreneur Stuary Skorman (who was also behind Reel.com, a mid-90s company that might have evolved into something like Netflix if he hadn’t sold it to Hollywood Video in 1998), clerk dogs taps the collective wisdom of 20 former video store clerks in an effort to create a more intuitive and more useful movie recommendation service.

The current beta of the service “works best” with crime and suspense movies, but already supports a wide variety of genres. Users can use a “Mash It” feature to adjust levels of various subjective qualities in a film to generate a new list: for instance, users might want more character depth and more complexity, but perhaps less disturbing imagery or suspense. When users find a movie they like, clerk dogs provides a direct link to get the movie from Amazon.com.

Clerk dogs is aiming to achieve a recommendation system that’s more akin to having a conversation with a serious film buff, rather than merely matching numbers and ratings from a (potentially vast) number of unknown users who, for all a user knows, may have panned a movie because it was physically damaged, or they didn’t like the case cover.

Editors' Recommendations