This week Netflix announced a winner in its contest to improve its movie rating software. A team of seven computer engineers, statisticians and machine learning experts, Bellkor’s Pragmatic Chaos, won $1 million for creating an algorithm that improves the current software by more than 10 percent.
Some context: Netflix wants happy customers who see the value from its movies-by-subscription service. If its customers love their movies more, they’re likely to remain customers longer, tell their friends to sign up, or spend more on the service.
There are two key takeaways here. First, by crowdsourcing the contest, Netflix gained access to the intelligence of a large community of experts. The prize is cheaper than the cost of developing the software in-house. Multiple teams from across the world participated, including a team of 20 that created the same result but demonstrated it just minutes after Bellkor’s Pragmatic Chaos. Darn that coffee break! Darn that crosstown commute!
Second, the teams all worked from the same data – the API, or application programming interface, was provided by Netflix. An API generally consists of a ball of data, coupled with rules about how to access and process it. This trend is increasing. A large part of Twitter’s growth is linked to its open API and the constant reinvention that accompanies it. For example, Twitscoop uses the same data as Twitter Search, but delivers a graphic interface, tag clouds and other enhancements.
The New York Times has taken steps to open its API, and The Guardian, an English newspaper, is also experimenting with the concept.
Kudos to these newspapers. Inside the API lies all the goodness of traditional journalism, which is desperately trying to establish a sustainable business model. Opening the API makes possible the kind of reinvention that we’ve seen with Google and Twitter.
Who will save newspapers? Perhaps it’s time to look past the usual suspects like journalists, investors and foundations. Let’s see what the programmers can do.