Posted by Ajay Kelkar on Mon, Jul 13, 2009
Netflix has a computerized recommendation system which seeks
to find the common threads in millions of people’s recommendations and throw up
“what’s best for you”? A lot of Online commerce sites have this kind of
recommendation engine at work and Amazon is another well known example.
Netflix had a unique challenge: they wanted to improve
Cinematch , their “recommendation
engine.” Cinematch is the bit of software embedded in the Netflix Web site that
analyzes each customer’s movie-viewing habits and recommends other movies that
the customer might enjoy.
Netflix wanted Cinematch to compete with the best minds “out
there” and anyone who showed a 10% improvement over Cinematch would stand to win $ 1 million!
Breaking
news:
On Friday, Netflix Prize team “BellKor’s Pragmatic Chaos” passed the mark,
qualifying for the $1,000,000 prize. The team includes engineers from AT&T.
Now, according to Netflix, Pragmatic Chaos — a coalition of employees from
AT&T, Commendo, Pragmatic Theory, and Yahoo Research — has figured out the
winning formula. Just last week , Pragmatic Theory submitted an engine that
showed an improvement of 9.78%, just a sliver off the 10% mark. Today, however,
with the 10.05% submission from Pragmatic Chaos, we could see this contest
finally rewarding the true genius of these scientists.
But they haven’t won yet. Their qualification triggers a
30-day count-down during which all teams have a final chance to improve their
efforts.
Clive Thompson has written this fascinating piece for The
New York Times
http://www.nytimes.com/2008/11/23/magazine/23Netflix-t.html?pagewanted=1&_r=2
Robert Grossman has this interesting piece on how companies
can learn from this regarding Analytics strategy.
http://blog.rgrossman.com/