Netflix Never Used Its $1 Million Algorithm Due To Engineering Costs

Netflix awarded a $1 million prize to a developer team in 2009 for an algorithm that increased the accuracy of the company’s recommendation engine by 10 percent. But it doesn’t use the million-dollar code, and has no plans to implement it in the future, Netflix announced on its blog Friday. The post goes on to explain why: […]

Netflix awarded a $1 million prize to a developer team in 2009 for an algorithm that increased the accuracy of the company's recommendation engine by 10 percent. But it doesn't use the million-dollar code, and has no plans to implement it in the future, Netflix announced on its blog Friday. The post goes on to explain why: a combination of too much engineering effort for the results, and a shift from movie recommendations to the "next level" of personalization caused by the transition of the business from mailed DVDs to video streaming.

Netflix notes that it does still use two algorithms from the team that won the first Progress Prize for an 8.43 percent improvement to the recommendation engine's root mean squared error (the full $1 million was awarded for a 10 percent improvement). But the increase in accuracy on the winning improvements "did not seem to justify the engineering effort needed to bring them into a production environment," the blog post said. By that time, the company had moved on anyway.

When Netflix announced the contest to improve the service in 2007, its business was centered on DVDs, which are dealt with by customers in periods of days or weeks and provide little granular data. Now that Netflix's primary offering is streaming, it has access to much more information: "streaming members are looking for something great to watch right now; they can sample a few videos before settling on one, they can consume several in one session, and we can observe viewing statistics such as whether a video was watched fully or only partially," reads the post.

This doesn't seem like much data, but it broadly affects the full screen of personalized recommendations a user might see on their Netflix home page. If a user begins a movie from "Imaginative Time Travel Movies from the 1980s" but quickly closes it, the homepage could shuffle that category down and place a new category at the prime top position that still speaks to the customer's watching history; something less sci-fi and more 1980s, or vice-versa.

(Mostly) gone are the days that customers would fill their DVD lists with artsy indie films or all of the Academy Award-winning documentaries they could, only for them to remain in queue purgatory. Netflix Watch Instantly is about the here and now, and Netflix is priming to respond to that time frame.