The $1 million question: How can one get inside the customer’s mind?

Netflix is betting $1 million that its subscribers can help the company enhance its e-commerce personalization techniques by coming up with a better way to recommend movies to its customers.

The bet is Netflix Prize, an award of $1 million to the first person or team that can help Netflix achieve certain accuracy goals in recommending movies based on customers’ personal preferences. To assist contestants in their efforts, Netflix has made available 100 million movie ratings logged by its subscribers; it has de-identified the ratings—which range from one to five stars—to keep subscribers’ data and opinions confidential.

Contestants will be provided with access to 100 million moving ratings collected from the company’s ratings information database and challenged to develop better software applications that will help Netflix deliver more accurate and personalized movie recommendations. "Recommendation systems covering a wide variety of categories will play an increasingly significant commercial role in the future," predicts Netflix co-founder, chairman and CEO Reed Hastings. "Right now we`re driving the Model T version of what is possible. We want to build a Ferrari, and establishing Netflix Prize is a first step."

The threshold required to win the Netflix Prize is a 10% improvement in accuracy over the current Netflix recommendation system. If no one wins the grand prize in 2006, Netflix will award a $50,000 progress prize to the contestant or team of contestants that makes the most significant advancement toward the goal. It will award a progress prize annually until someone wins the grand prize.

Netflix, No. 21 in the Internet Retailer Top 500 Guide to Retail Web Sites, currently uses a proprietary technology, Cinematch, to predict whether someone will enjoy a movie based on how much they liked or disliked other movies.