Copyright ©2007 NPR. For personal, noncommercial use only. See Terms of Use. For other uses, prior permission required.

RENEE MONTAGNE, host:

On Mondays, we talk about technology. And today the online rental company Netflix is trying to improve its recommended movies. A year ago, Netflix offered $1 million to anyone who could improve its high-technology movie referral system by 10 percent. No one's met that target quite yet, but today Netflix will award a $50,000 progress prize to the team that came the closest.

NPR's Scott Horsley takes a look.

SCOTT HORSLEY: When it comes to renting movies, Kensington Video in San Diego does it the old-fashioned way. There's not a computer in sight. Customers looking for recommendations turned to real people like Pam Cisneros.

Ms. PAM CISNEROS (Owner, Kensington Video): You know, I have a little lady coming earlier today, and she wanted another really good recommendation. Well, I know she doesn't like something that has lots of sex in it and something that's too brutal, as far as the violence goes - just a good movie without all that had extra stuff in there. They do make a few of those still.

HORSLEY: Netflix can't exactly duplicate that kind of personal attention with seven million subscribers, but the online rental company does provide computerized recommendations. Customers rate movies on a scale of one to five stars.

Vice President Jim Bennett says the company's track record is pretty good.

Mr. JAMES BENNETT (Vice President, Netflix): We get, basically, three out four correct, within a half of a star - which is pretty high, actually.

HORSLEY: Netflix uses what computer scientists call a neighborhood approach.

Mr. BENNETT: So if you rated "Gone with the Wind" and you rated "Anna Karenina" both positively, and we find many people who have done that, then we're able to find a correlation between those two, so that if somebody comes along and rates "Gone with the Wind" but hasn't told us anything about "Anna Karenina," we should adjust their prediction for "Anna Karenina" positively.

HORSLEY: But the Netflix system only goes so far. Thousands of statisticians and computer scientists have been competing to improve it. No one scored the 10 percent improvement needed to claim the million dollar prize.

But a team from AT&T Labs came close. By slicing and dicing millions of Netflix's ratings, the team was able to find patterns that seemed to suggest common factors that might lead certain customers to like certain films. Those factors could be dialogue or action or popular stars, but the team doesn't pretend to know.

Statistician Bob Bell says he's not even much a film buff. What interests him is the data.

Dr. BOB BELL (Statistician, AT&T Labs): As far as we were concerned, the movies were just an ID number.

HORSLEY: If that seems a little too mathematical, you might want to try another movie recommendation site: whattorent.com. Adam Geitgey says that site asked would-be runners a series of personal questions, sort of like a matchmaker about to arrange a blind date.

Mr. ADAM GEITGEY (Founder, Whattorent.com): People relate to movies like they relate to people, so then if we can figure out what kind of person you are and what kind of person the movie would be, we could match you with a movie.

HORSLEY: But that's not really practical on a large scale. Netflix Vice President Bennett says the beauty of his company's algorithm them is it can work automatically, and with any kind of product. That's why not only Netflix, but all kinds of e-commerce sites are anxious to see it work better.

Mr. BENNETT: It's our "Who Wants to be a Millionaire for Geeks." We're looking forward to giving out that million-dollar prize and getting that a little bit of extra accuracy that could make all the difference for our business.

HORSLEY: And recommendation algorithms aren't just good for companies that want to move more merchandise. Computer scientist John Riedl of the University of Minnesota argues they're also good for customers looking to find that out-of-the-way film that isn't a blockbuster hit.

Dr. JOHN RIEDL (Computer Scientist, University of Minnesota): What I'm excited about in recommenders is the possibility for letting the little hit matter, right? For letting all of us who have our own quirky, special little movie or book taste be recommenders for the people who like the kinds of things that we recommend.

HORSLEY: If Riedl's right, moviegoers and other online merchants could be giving Netflix's new and improved algorithm millions of thumbs up.

Scott Horsley, NPR News.

Copyright © 2007 NPR. All rights reserved. No quotes from the materials contained herein may be used in any media without attribution to NPR. This transcript is provided for personal, noncommercial use only, pursuant to our Terms of Use. Any other use requires NPR's prior permission. Visit our permissions page for further information.

NPR transcripts are created on a rush deadline by a contractor for NPR, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of NPR's programming is the audio.

Comments

 

Please keep your community civil. All comments must follow the NPR.org Community rules and terms of use, and will be moderated prior to posting. NPR reserves the right to use the comments we receive, in whole or in part, and to use the commenter's name and location, in any medium. See also the Terms of Use, Privacy Policy and Community FAQ.