The NETFLIX Algorithm Contest: A Winner Emerges?

Netflix, the web based DVD rental service, launched as contest in 2006 offering a 1 million dollar prize to team that develops a recommendation algorithm that is shown to be 10% better that Netflix’s current recommendation engine. To those in the programming community, the challenge has been compared to scaling Mount Everest. Has the summit been reached?


It seems the two contest front runners, Team Pragmatic Theory and Team Bellkor in BigChaos, have joined forces and submitted an algorithm that was 10.05 percent better than the one Netflix uses to recommend movies to its subscribers. The result was published on the Netflix Prize leader board on June 26, 2009.

From a promotion law point of view, the contest winner must grant to Netflix, an irrevocable, royalty free, worldwide non-exclusive license under the contest entrant’s copyrights, patents or other intellectual property rights in the winning algorithm. In addition, a description of the algorithm, but not the source code, will be published on the Netflix site. I.e., the winners will “describe to the world how [they] did it and why it works.” Like all well conceived contests, the official rules provide a mechanism for determining the contest entrants actually created the entry submitted. It appears however, that the winners are not prohibited patenting their entry and winners are not prohibited from charging others for the use the algorithm.

This contest, where competitive incentives are offered as an alternative to in-house research and development, looks like a new, workable model to foster innovation.



Permanent link to this article: