clock menu more-arrow no yes

Filed under:

Recommendation is one of the biggest issues facing streamers like Netflix, HBO Max, and more

Directors, engineers, and executives acknowledge the issue

If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.

The Farewell director Lulu Wang said something during a recent roundtable with The Hollywood Reporter that neatly sums up the biggest problem facing the streaming revolution: there’s too much stuff.

Wang had been offered more money by a major streaming platform for her film, but opted to go with A24, a smaller, more traditional studio. Director Todd Phillips said a filmmaker “like Lulu would get lost on Netflix,” and Martin Scorsese added that there are “too many things.” Wang compared big streaming services to Spotify and iTunes where there’s a never-ending list of music to discover, but in the end “if people don’t know how to find you, it doesn’t matter how wide your audience is.”

“One thing we sometimes don’t talk about with these bigger streaming platforms is that it’s a different business model,” Wang said. “They’re building their brand, and when you are an established filmmaker, you are a brand that they want to partner with to help build their own brand. But with newer filmmakers, newer voices, you don’t have a brand.”

Netflix is sometimes regarded as a necessary alternative for filmmakers, especially with blockbusters from Disney and Warner Bros. dominating the box office. Ted Sarandos, the company’s chief content officer, recently spoke at an Aspen Institute summit and declared that Netflix isn’t hurting cinema, but saving it. The streaming service is taking chances on movies that other studios won’t, and producing more than any other company. Netflix paid north of $160 million for Scorsese’s The Irishman nearly a decade after the director tried to get it made in the regular studio system. The streamer released nearly 60 original English-speaking films in 2019 — about triple what Disney produced.

More is crucial to Netflix’s strategy. It’s the streamer that has something new to offer every week. Industry observers have argued that floods of content are burning out consumers, with FX president John Landgraf adding that binge options from streamers that drop seasons all at once mean shows leave public conversation after two weeks. Of course, Netflix’s product team disagrees.

“These myths that the competition puts out there, and the media fuels, are absurd,” Todd Yellin, Netflix’s vice president of product, told Vulture in June 2018. “It’s about getting that title to the right person. If we plastered the whole service and everyone got Bright, what would happen to that documentary? We are able to unlock audiences for titles that you wouldn’t expect.”

Yellin’s point may be true, but there’s an understood caveat that Bright would take the main spot on most people’s homepages, while documentaries might slide down a couple of rows. People have to find one while being presented with the other. Netflix runs about 400 A/B tests a year, mostly focused on recommendation algorithm experiments. The goal is to ensure that people who watch a specific type of content will receive recommendations on their homepage for similar shows and movies. That doesn’t always turn out to be true, though. Take Kenya Barris’ Astronomy Club, an all-black sketch comedy show that debuted in early December.

Buzzfeed’s Michael Blackmon wrote that Astronomy Club “for one reason or another, got lost in the madness of the concluding decade.” A colleague who searched for the show on Netflix was told by the streamer that Astronomy Club was an 85 percent match based on other titles she’s seen, but it never populated on her homepage. Astronomy Club isn’t just a random show, either. Black-ish creator Kenya Barris produced the series following his reported $100 million deal with the streamer. Since Netflix doesn’t release numbers, it’s unclear how many people watched the show.

Search and discovery is one of the biggest issues other companies coming into the space are thinking about. HBO Max is rolling out an entire marketing campaign around using human curators instead of relying on recommendation algorithms. WarnerMedia chief technology officer Jeremy Legg spoke to Cheddar about the company’s feelings around search heading into Max development.

“It’s been tried through algorithms in order to provide those recommendations to consumers,” Legg said. “There’s been some success with that. There’s some balance between curation and algorithms, especially with search and discovery for consumers, but I don’t think the sweet spot’s been hit yet.”

Other streaming services like Quibi are looking at search and discovery through platform-specific behavior. Executives see the problem with a setup like Netflix has as too time-consuming. It can take 10 minutes to find a show. The whole purpose of Quibi is to provide instant entertainment for short periods of time. The company will use machine learning to find what subscribers want to watch and surface the closest recommendation in order to ensure people aren’t just scrolling when they open the app.

People are being directly served more content than ever before. There are immense positives to this: more jobs for people working in the industry, more options for people to watch, and chances for new artists to break through. There are some consequences, however, and conversations these past few weeks have shown how search and discovery needs to improve. Look at a show like Tuca and Bertie. After it was canceled following its beloved first season, show creator Lisa Hanawalt said the recommendation algorithm was to blame.

The reason The Farewell’s Lulu Wang ultimately signed with A24 for a much smaller paycheck than a major streamer had to do with visibility. A24 ensured that as many people as possible knew about the movie and went to see it, keeping it in theaters for four months on top of festival runs. On a streaming platform, Wang argued, it’s all up to chance.