Dorothy Crenshaw
July 28, 2009 | 03:19:57
Partly because I gave four stars to Woody Allen’s Manhattan, Netflix is recommending the dark and brooding East German indie The Lives of Others. Hmmm.
I don’t know about you, but for me, recommendation engines that try to predict our likes and dislikes in books and movies usually miss the mark. I’m talking about the collaborative filtering tools you find on sites like Netflix and Amazon.com. I’ve always chalked it up to the quirks of personal taste, however. Normally I hate violent movies, yet Kill Bill 2 is on my Top 5 favorites list. Who knows why…I defy an algorithm to figure out that one.
Which is why I was interested in the attention around the Netflix Prize. That’s the global competition that promises to award $1 million to the person or team who can improve on Netflix’s current algorithm for predicting member preferences by at least 10 percent.
Although some have dismissed it as a PR stunt, it’s designed as a real research project, which seems to have generated lots of positive PR for the brand. And there are a couple of counterintuitive aspects to the Prize that make it worth watching.
First, it’s an admission by Netflix that its internal efforts to improve the preference engine simply haven’t worked. It needs help. Normally that might reflect badly on the brand, the service, and its technology capabilities, but it hasn’t. Netflix seems to have tapped into the crowdsourcing thing at just the right moment.
I also like the packaging of the event, down to its grandiose label. No cheesy “contest” here. It’s a “prize,” as in “Nobel.” This isn’t just another Ben and Jerry ice cream flavor, folks! We’re talking innovation.
Finally, it has real drama. The competition culminated in an exciting, horserace-style photo finish at Sunday’s deadline. Though the actual winner won’t be announced until September, you’re able to follow the rankings on a leaderboard, which showed a last-minute surge by an upstart team, complete with cheers, jeers, and impressive participation by over 44,000 valid entrants. The blogosphere has covered it heavily as well.
In its longevity and substance, the Prize conveys a true commitment by Netflix to both technology innovation, or at least enhancement, and customer service. And there are real learnings here. Today’s New York Times story about the value of teamwork to an undertaking like this actually positions the Prize as precedent-setting for predictive modeling.
It’s an impressive case history on how to run a contest…oops, global competition. I’m curious to see how the winning entry will be conveyed to everyday, non-techie Netflix members like me. But, the real proof, of course, will be in how it performs in picking winners for movie night.
By the way, I did see The Lives of Others. Rent it now. Amazing movie.