I’ve been a Netflix customer since March 2002 and think the service is fabulous. But, the one thing that’s not so great is their movie recommendations. So, I was thrilled when in October 2006 Netflix announced their $1 million prize for the creation of a movie recommendation algorithm that could best their current system by 10%.
Based on an article in the March 2008 issue of Wired, progress has slowed but there have been over 25,000 entrants. The story profiles an interesting newcomer to the race who is adding psychology to the mix of mathematics.
There’s so much right about this contest…
- Demonstrating, with a big bottom line number of a million bucks, that they’re dedicated to improving their product
- Contributing to the community of mathematicians, statisticians and others by providing a venue to share knowledge
- Generating press coverage for their company in a way that supports the very essence of their brand
- Engaging their customers by posting progress on a leaderboard on their site (an entrant as of 3/16/08 has reached 8.93% improvement)
In contrast, one site that consistently makes great recommendations and has introduced me to a number of artists is Pandora Radio. Put in the music you like and it will create a “station” of music based on your tastes. According to their website, they “capture the essence of music at the most fundamental level…everything from melody, harmony and rhythm, to instrumentation, orchestration, arrangement, lyrics, and of course the rich world of singing and vocal harmony. It's not about what a band looks like, or what genre they supposedly belong to, or about who buys their records - it's about what each individual song sounds like.”
Hmmm…maybe if they could do the same thing for movies…
