Netflix Gives Out $1 Million for Movie Research Winners
A Netflix contest to find out who could produce a better recommendation system has made one research team $1 million richer, but the real winner is Netlix, which is receiving premium research for pennies on the dollar.
Netflix, the online DVD rental company that has set in motion a chain of events that is close to burying rival Blockbuster Video, has just finished up a three-year process of helping customers select movies. To help their customers, Netflix has overseen a contest that pitted researchers against one another to come up with a better system of suggesting movies. The contest prize of $1 million has just been awarded to a seven-member research group, but Netflix has seen such good results that another contest for a $1 million is in the works.
The contest, designed to help the company offers suggestions, is really an ingenious play by the company, which appears to be receiving several million dollars worth of research for the relatively paltry sum of just one million. Launched in 2006, Netflix began implementing ideas from contestants just three weeks after receiving suggestions. BellKor’s Pragmatic Chaos team won the contest. Objective measures showed that the team’s methodology led to an improvement in recommendations of 10%, but that this narrow improvement actually led to picks that are twice as good as the current model.
The importance of the recommendation system for Netflix is about customer service to some extent, but is primarily a gambit to retain customers. Those who watch a lot of movies and television on DVD will eventually exhaust their movie "queu" and consider canceling their accounts. A good recommendation system, however, points such users in the direction of more DVDs to rent, thus keeping customer squarely in the Netflix system. Said Netflix CEO Reed Hastings, "I was stunned at how the Netflix Prize created its own economy of researchers competing and collaborating."
The contest, designed to help the company offers suggestions, is really an ingenious play by the company, which appears to be receiving several million dollars worth of research for the relatively paltry sum of just one million. Launched in 2006, Netflix began implementing ideas from contestants just three weeks after receiving suggestions. BellKor’s Pragmatic Chaos team won the contest. Objective measures showed that the team’s methodology led to an improvement in recommendations of 10%, but that this narrow improvement actually led to picks that are twice as good as the current model.
The importance of the recommendation system for Netflix is about customer service to some extent, but is primarily a gambit to retain customers. Those who watch a lot of movies and television on DVD will eventually exhaust their movie "queu" and consider canceling their accounts. A good recommendation system, however, points such users in the direction of more DVDs to rent, thus keeping customer squarely in the Netflix system. Said Netflix CEO Reed Hastings, "I was stunned at how the Netflix Prize created its own economy of researchers competing and collaborating."

Use the feedback form below to submit your comments.

Use the form below to email this article to your friends.

- Netflix Throttling Has Some Subscribers Seeing Red
- Be ready for the new gaming experiences
- Songwriting Contest Goes Global
- Barron Hilton Pleads No Contest
- Ray Liotta Pleads No Contest
- Paris Hilton Pleads No Contest
- DJs Sacked Over Fatal Water Drinking Contest
- Haley Joel Osment Pleads No Contest
- Song Contest Offers Something For Everyone
- Eurovision song contest



