Netflix, Inc. (NFLX) Algorithms: What People Want To Watch?

Updated on

Netflix, Inc. (NASDAQ:NFLX) queues reflect what people say they want to watch, but anyone who uses the service knows it’s often easier to choose from among the recommendations Netflix, Inc. (NASDAQ:NFLX) sends your way. It turns out that Netflix engineers are well aware of the difference between the movies you say you will watch and the ones that actually appeal to you, and they work hard to put the second group front and center, reports Tom Vanderbilt at Wired.


In the past, Netflix, Inc. (NASDAQ:NFLX) used people’s ratings to make recommendations. It wasn’t just a matter of finding people who had liked a few movies in common, the entire spectrum of ratings from good to bad formed a sort of ratings fingerprint that could be used to match two users as having similar tastes. But that approach has two major problems, one involving the switch to streaming video, and the other simple matter of pride.

“When we were a DVD-by-mail company and people gave us a rating, they were expressing a thought process,” says Netflix, Inc. (NASDAQ:NFLX) engineering director Xavier Amatriain. “You added something to your queue because you wanted to watch it a few days later… With instant streaming, you start playing something, you don’t like it, you just switch. Users don’t really perceive the benefit of giving explicit feedback, so they invest less effort.”

Netflix found feedback doesn’t match what people want to watch

But even when people do provide explicit feedback, Netflix, Inc. (NASDAQ:NFLX) has found that it doesn’t match what they actually want to watch.

“People rate movies like Schindler’s List high, as opposed to one of the silly comedies I watch… if you give users recommendations that are all four- or five-star videos, that doesn’t mean they’ll actually want to watch that video on a Wednesday night after a long day at work, explains Carlos Gomez-Uribe, Netflix VP of product innovation and personalization algorithms.

User data far more important than ratings

Amatriain agrees that user data is far more important than ratings, because it has proven to be an accurate predictor of what people will select. “We know that many of the ratings are aspirational rather than reflecting your daily activity,” he says. Ratings don’t tell you how likely a person is to watch a given movie, they tell you some combination of perceived quality, importance, as well as information about how a person sees him or herself, but Netflix, Inc. (NASDAQ:NFLX) is trying to cut through all of that and find a movie you actually want to watch.

“A lot of people tell us they often watch foreign movies or documentaries. But in practice, that doesn’t happen very much,” says Gomez-Uribe.

Leave a Comment