I have been following Netflix unique experiment to improve its Web site’s movie recommendation system. This week Netflix announced the winner of a three year contest with the winner BellKore comprising of statisticians, computer scientists, data mining experts netting a cool million dollars.
The rules of the competition was fairly straightforward. The qualification for the prize was that the winning team has to improve by at least 10% the prediction of what movies customers would like as measured against the actual ratings. The teams were grappling with a huge data set of more than 100 million movie ratings.
Over the past three years there have been 44,014 entries from 5,169 teams in 186 countries vying for the top prize
I think with this experiment and with Google’s experiment with crowdsourcing described here, there will be a significant shift towards innovation management. The fact that there exists more intelligence and wisdom and the collective effort outside the company’s eco-system has gained credibility. I expect many such organizations embarking on the contest mode to solve intractable problems.
There are a number of lessons that this contest brings about.
First, it indicates that there can be a marketplace for innovation where companies could post their product development challenges and for an interesting contest, the best brains are willing to compete. It sharpens their own abilities.
Second as the BellKore team and other teams demonstrated there is a willingness for disparate people to actively collaborate. While cooperation and collaboration within many organizations has been challenging, I wonder how such disparate people could come together and collaborate easily for a bigger goal.
Third, for people who believed in having an inhouse R&D and saw that as a competitive advantage, this experiment seeks to blow that myth away.
Note: Netflix Prize 2 would challenge competitors to recommend movies based on demographic and behavioral data.