Netflix’s Xavier Amatriain and Justin Basilico provide the history:
In 2006 we announced the Netflix Prize, a machine learning and data mining competition for movie rating prediction. We offered $1 million to whoever improved the accuracy of our existing system called Cinematch by 10%….A year into the competition, the Korbell team won the first Progress Prize.
Amatriaan and Basilico go on to explain why Netflix will never use the code that won the prize. Basically, technology, and consumer preferences, progress, and “did not seem to justify the engineering effort.” “It wasn’t just that the improvement was marginal,” said Mike Masnick, “but that Netflix’s business had shifted and the way customers used its product, and the kinds of recommendations the company had done, had shifted too.”
The Netflix Prize was quite different from most philanthropic contests. Still, I wonder if we might discern lessons for those running contests today: