So, three-quarters of a year into Netflix’s million-dollar movie-prediction challenge, it seems that the leading team is three-quarters of the way towards the goal of improving Netflix’s recommendation system by 10%.
Which makes me think about the role of prizes in innovation. Clearly, they are starting to regain some currency amongst the technorati. The recent Netsquared conference-cum-contest seems to me a like a prime example of the trend. Re-reading the initial article about the Netflix challenge was interesting — challenges clearly have a lot of untapped power, but several conditions have to be in place for them to be effective:
1) The challenge has to have a specific, measurable, unambiguous, non-subjective goal. In Netflix’s case, the challenge is to develop an algorithm that is 10% more accurate than Netflix’s current system. The X Prize awarded $10 million to “a spacecraft capable of carrying three people to 100 kilometers above the earth’s surface, twice within two weeks.”
2) Prizes don’t operate in a vacuum; they leverage existing resources. Capturing a prize costs lots of money — in the X Prize’s case, the teams spent nearly $100 million on their projects. The leading Netflix challenge team is a group of four people working full time. That means there needs to be other sources of philanthropic money available to subsidize the actual work.
Food for thought for nonprofit technology folks thinking about contests and prizes.