How the hits keep on comin’ for Netflix
Service helps shows get discovered using viewer data
NEW YORK — Netflix wants subscribers to know it’s looking out for them.
For instance, the average Netflix subscriber might never guess that its dark superhero drama Jessica Jones might strike similar chords as the zany hijinks of Unbreakable Kimmy Schmidt. Netflix is happy to help you make the connection.
Much of the attention showered on this streaming-video giant in recent years has dwelled on its insatiable appetite for original content and for creators to produce it.
“We want to appeal to as many different people as possible and appeal to the many moods that each person has,” says Todd Yellin, Netflix vice-president of product innovation.
But, this service’s multibillion-dollar annual outlay for new programming necessitates another challenge: helping each program get discovered by the subscribers most likely to enjoy it. Four out of five of the shows watched on Netflix were found by its subscribers thanks to recommendations offered to them, Netflix says.
Those suggested new favourites are much more customized for each subscriber than might be evident from a glance at the Netflix home page.
Most every row of program suggestions (even generic-seeming categories like comedies and dramas) is tailored for each subscriber, Yellin says. And, how the rows are arranged vertically on the home page is a function of the subscriber’s demonstrated genre preferences.
“You might have comedies as your fifth row,” says Yellin, “and for another person it might be 25th. And, someone else might not get a comedy row at all.”
So, your Netflix is different from everybody else’s. But, where do these tips come from?
“It’s very important that the titles most relevant to each person bubble up to the top of the catalogue,” says Yellin. “And we want those relevant titles to be diverse. We don’t want to make the amateur mistake of getting caught in an echo chamber, such as just because you watched one horror title, slapping in front of you nothing but more horror titles.”
Yellin likens the process of providing bespoke TV for each customer to a three-way collaboration.
First, a legion of Netflix “taggers” screens every program, tagging different elements that compose it. This data is crunched and continuously refined by the company’s secret-sauce algorithm. And then viewer habits gathered by Netflix from its 100 million accounts worldwide add more grist to the mill. Netflix can then take a “gateway” program and point the person watching it to other unexpected or unknown fare with presumably similar appeal.
Consider Ozark, a recently released original drama series starring Jason Bateman as a money-laundering family man who’s seriously jammed up with the Mexican drug cartel he works for.
“We’ve found that people who tend to watch Blacklist and House of Cards tend to like Ozark,” says Yellin. “But another kind of person who will find he likes Ozark is a fan of Narcos and El Chapo and other drug-cartel-oriented dramas and documentaries.”
But wait, there’s yet another “taste community” rallying to Ozark, says Yellin: fans of the 2015 film The Big Short.
“It’s not like we could have guessed this ahead of time,” says Yellin. “We just track which shows tend to cluster together. Who would have thought that Jessica Jones and Kimmy Schmidt would cluster together?”