Netflix Prize the Sequel: Even Better Movie Recommendations?

netflix prize

The Netflix Prize was announced three years and has just ended. The goal of the contest is to improve movie recommendations for Netflix customers (and presumably increase Netflix revenue and profit). Participants were given a large data set of “100 million ratings from over 480 thousand randomly-chosen, anonymous customers on nearly 18 thousand movie titles” (read more about the original contest in the rules). The winner (and there are two possible ones) will be announced in September along with a new contest.

netflix prize leaderboard

Yesterday Chief Product Officer Neil Hunt made this announcement:

The advances spurred by the Netflix Prize have so impressed us that we’re planning Netflix Prize 2, a new big money contest with some new twists.

Here’s one: three years was a long time to compete in Prize 1, so the next contest will be a shorter time limited race, with grand prizes for the best results at 6 and 18 months.

The first Netflix Prize seems to have been an obvious success for Netflix. They got tons of PR, a lot of geek goodwill and in the end a better algorithm. However, no one but Netflix gained from the work contestants created. In a world where algorithms make or break a product it’s no surprise that the company wants to keep the work secret. However it’s really great to see the next line of the announcement.



While the first contest has been remarkable, we think Netflix Prize 2 will be more challenging, more fun, and even more useful to the field.

This could mean any number of things. However, my hope is that it leads to an opensourcing of any code and algorithms created during the Netflix Prize 2 contest. It seems likely that it could be done after the contest and that the work could really benefit the field. If this is the plan then either Netflix has changed their priorities around secrecy or the contest is not around such a high impact area of the site (or I am totally wrong). Time will tell, but it’s great to see Netflix sponsoring this type of innovation in the wild.

  • http://twitter.com/mcobrien Michael O'Brien

    I had thought the winning algorithm was supposed to be made generally available. From the rules page you linked to:

    “To win and take home either prize, your qualifying submissions must have the largest accuracy improvement verified by the Contest judges, you must share your method with (and non-exclusively license it to) Netflix, and you must describe to the world how you did it and why it works.”

  • Voice

    Why should Netflix spend substantial amounts of money to pay for these algorithmic advancements, only to then give away their competitive advantage by open sourcing it ?

    Some people just don’t live in the real world.

    The fact that it took a cash prize to incentivise these improvements says to me that the open source movement can kid themselves as long as they want, but the real advances will only come when someone is paying the bills.

  • http://cowsandmilk.net David Hall

    The teams aren’t required to open source the product. You’re just required to share the knowledge. This likely will go through a patent, then publication in the literature.

    I’m unsure why Brady thinks that Netflix wants to keep the work secret since the rule that the knowledge must be shared has been known from the start. And much of the knowledge out of this has already been shared. It’s widely celebrated the collaboration between teams in the effort to win.

    Additionally, a quick Google Scholar search shows the large number of conference proceedings related to work on the Netflix Prize.

  • http://www.sskbank.com ssk sorgulama

    “To win and take home either prize, your qualifying submissions must have the largest accuracy improvement verified by the Contest judges, you must share your method with (and non-exclusively license it to) Netflix, and you must describe to the world how you did it and why it works.”

  • http://kallistec.com/ Dan DeLeo

    Github has a recommendation contest going on right now. One of the rules is that your code needs to be open sourced after the contest. On the other hand, the prize, while highly desirable, is much less than a million dollars.

    http://contest.github.com

  • http://lordbron.wordpress.com Thomas Ortega II

    Uh, Brady, I think you missed that it is public like David Hall said.

    @voice, I think you miss the point.

    Here’s my take on it:
    http://lordbron.wordpress.com/2009/07/02/the-future-of-biz-crowdsourcing/

    The whole point is to keep the innovation going, not horde the “secrets”