From its humble beginnings as a mail-order service for physical storage media, Netflix has grown into a monolithic entertainment giant. As of October 2018, the company reported 130 million paid subscribers worldwide, with a forecast of millions more being added every quarter.
That growth depends in large part on making sure people are finding programming options and then binge-watching them until work and relationships begin to suffer. With so much on the line, it’s no wonder Netflix has a strategy for keeping viewers tuned in—one that has a lot in common with book covers, according to Thrillist.
In a recent Medium post, Netflix programmers revealed that their algorithm for customizing recommendations comes down to more than just a laundry list of titles. Each curated selection will use a still image—which the service calls a landing card—that best represents your interests based on your viewing history.
As a hypothetical example, a menu option for 1993’s Jurassic Park might show off the dinosaurs if you have a tendency to watch creature features. If your queue is full of Jeff Goldblum movies, then you might instead see a shot of the actor (though probably not that shirtless one). In the case of 1997’s Good Will Hunting, a love for dramatic movies might have Netflix show you a sobbing Matt Damon. If you watch a lot of comedies, they’ll show you Robin Williams instead.
Artwork makes up a big part of how Netflix viewers make selections, with one 2016 company survey indicating that the images made up 82 percent of their focus while browsing. A movie or series might have up to a dozen different images, with your viewing habits ranking which one might be most likely to get you to click on it.
It’s the viewer’s hierarchy, not Hollywood’s, that informs those choices. If you’ve mainlined several Jack Black movies, then you might see his face instead of John Cusack’s in a box for 1999’s High Fidelity—even though Cusack is the star and Black has a supporting role. For cases in which stars might have equal appeal, the user’s preferences will help determine what they see. Uma Thurman enthusiasts will notice her in art for 1994’s Pulp Fiction; John Travolta fans will get his image on their screens.
While this is an effective strategy, some critics of the approach say that this level of customization removes the discovery aspect of entertainment consumption. If Netflix is feeding you a diet of things you already like, how will you discover anything new and different that you may also like? The answer to that probably lies in the very analog strategy of asking a friend.