I’m sure many of you have seen this story by now… Netflix (the DVD by mail company) wanted to improve its relevance ranking algorithm by 10% (no small feat). With some improved math, Netflix customers would get more accurate predictions for the movies they are most likely to enjoy and potentially learn about more niche content (and, of course, with a better algorithm Netflix secures more rentals and happier customers).
Every company that relies on relevancy algorithms is constantly tweaking their recommendation software to improve results (Google, Amazon, etc.). What’s interesting in the Netflix case is that they opened up their data and invited anyone to take a crack at it. The team that could improve the ranking by the required 10% would get $1 million dollars. Last week the prize was finally awarded (after over 2 years of teams trying to crack the barrier). Do you see the significance? Netflix gets a solution that works (and a little publicity in the process), without having to devote significant company resources (besides the million I guess). Welcome to the world of crowdsourcing – getting the crowd to do your work for you.
You can read more on this story from the NYT BITS blog.
