Why Netflix Keeps Accidentally Burying the Best Movies on Its Own Platform

Netflix has a genuinely strange problem. It is simultaneously the most powerful distribution platform in the history of cinema and one of the least effective places to discover a movie worth caring about. It contains multitudes, as they say — specifically, it contains Okja, The Irishman, Roma, and Beasts of No Nation, all buried somewhere beneath seventeen seasons of reality competition content and a carousel of thumbnails that have been A/B tested into meaninglessness.

This is not an accident. It is, in the most frustrating possible way, a feature.

The Algorithm Has Opinions, and They're Boring

Here's the core issue with how Netflix (and Disney+, and Max, and most of the major streamers) thinks about film discovery: the recommendation engine is optimized for watch-time completion, not for the quality of the experience. A movie that holds 80% of its audience through the credits is a better algorithmic performer than a movie that holds 40% of its audience but leaves them thinking about it for three weeks.

Formally experimental films — the ones with unusual structures, challenging runtimes, deliberately slow pacing, or narratives that refuse to explain themselves — are algorithmic poison. They generate early dropoffs. Early dropoffs tell the system the movie is bad. The system stops recommending the movie. The movie disappears into the catalog, technically available but functionally invisible.

This is how Netflix managed to acquire Bong Joon-ho's Okja in 2017, generate significant controversy about its Cannes premiere, and then essentially let it vanish from cultural conversation within six months. The movie is still there. Go find it. I'll wait. It took you a minute, didn't it?

The Thumbnail Is the Message

There's a specific visual language that major streaming platforms have developed for their thumbnails, and it is the enemy of interesting cinema. The formula is roughly: one or two characters, facing the camera or slightly away, expression hovering between determined and concerned, title in a clean sans-serif font. This works great for thrillers and procedurals. It is completely useless for communicating anything about a film like First Reformed or Zola or The Lighthouse.

Marketing a formally unconventional film on a major platform requires convincing an algorithm that the film belongs in a category the algorithm already understands. Which means the marketing team has to make The Lighthouse look like a horror movie, or make First Reformed look like a religious thriller, or make Zola look like a prestige drama. None of these framings are wrong, exactly. But they all attract the wrong audience, who then bounce early, which tanks the algorithm performance, which buries the movie.

It's a self-defeating cycle that the platforms have shown almost no interest in breaking, because breaking it would require accepting that some films are valuable even when they don't perform well by completion-rate metrics.

Meanwhile, on Mubi

While Netflix is busy not recommending its own interesting catalog, boutique platforms have quietly built the most exciting film discovery infrastructure in American streaming history.

MUBI in particular has figured out something that the big streamers haven't: curation is a product. The platform's rotating library of 30 handpicked films — plus its expanding catalog of permanent titles — functions more like a well-programmed art house cinema than a content warehouse. You show up not knowing what you want, and the platform has a point of view about what you should watch. That point of view is developed by humans with taste, not by an algorithm trying to minimize churn.

The Criterion Channel operates on similar logic. It's not trying to be everything to everyone. It's trying to be exactly one thing to a specific audience: the best possible archive of cinema history, organized with genuine intellectual care. The result is a platform where discovery feels like discovery — where you might sit down to watch a Cassavetes film and end up three hours later having stumbled into Agnes Varda.

Kanopy, available free through most US public libraries, is doing something even more radical: making genuinely challenging cinema available at zero cost to people who might never have encountered it otherwise. The audience for Tangerine on Kanopy is different from the audience that sought it out at Sundance, and that's a good thing.

The Disney+ Problem Is Actually Worse

At least Netflix occasionally acquires something interesting. Disney+ has essentially committed to never programming anything that a twelve-year-old would find confusing, which is a coherent business strategy but a disaster for cinema as a medium. The platform's theatrical acquisition policy is almost entirely franchise-adjacent, and its algorithm is built around the assumption that if you watched one Marvel movie, you probably want to watch all of them, forever, in sequence.

This isn't a criticism of Marvel movies. It's a criticism of a platform that has decided its audience is a single, unified entity with identical tastes, rather than a collection of individuals who might want different things on different nights.

What's Actually Getting Discovered

Here's the genuinely weird thing: unconventional films are finding audiences. They're just not finding them on Netflix. They're finding them through film Twitter, through Letterboxd lists, through podcast recommendations, through the kind of word-of-mouth that happens when someone watches something genuinely strange and needs to tell everyone they know about it immediately.

The discovery infrastructure for interesting cinema has essentially moved off the major platforms entirely. People find out about a film somewhere else, then go to MUBI or the Criterion Channel or their local library's Kanopy account to actually watch it. Netflix is increasingly just where those same people go to turn their brains off afterward, which is fine, but it's a long way from the promise of being the future of film.

The algorithm isn't going to fix itself. The boutique platforms aren't going away. And the movies worth watching are still out there, just not where the biggest companies want you to look.