Streaming Users Consume More Diverse Content

In 2018, a French survey asked thousands of adults about their cultural consumption habits. When researchers Samuel Coavoux and Théo Aussant analyzed the data years later, they found something that contradicted what many critics had been saying about streaming platforms: people who used Spotify, Netflix, and similar services consumed more diverse content than those who didn't. Not less. The filter bubble thesis—the idea that algorithms trap us in echo chambers of our own preferences—didn't hold up. At least not in the simple way we'd imagined.

The Diversity Paradox

The numbers tell a more complicated story than either streaming evangelists or critics typically acknowledge. Streaming platforms do increase the diversity of what people watch and listen to, but the effect varies wildly by medium. Music streaming shows small gains in variety. Movies show moderate improvements. Television shows the biggest jump—people who stream TV consume substantially more diverse content than those relying on traditional broadcast or cable.

Why the difference? The answer lies less in how algorithms work and more in how streaming reconfigures the market itself. Netflix, Disney+, and Prime Video don't just distribute shows—they produce them, fundamentally changing what gets made and for whom. Music streaming, by contrast, mostly reshuffles existing catalogs. The platforms have virtually unlimited shelf space and every incentive to cater to niche tastes, but they can only recommend what already exists.

This explains why streaming's impact on cultural diversity comes primarily through market transformation rather than through changing how individual consumers experience culture. The algorithm isn't magic. It's infrastructure.

How Platforms Actually Learn

Streaming services track behavior with unsettling precision. Every pause, every skip, every replay gets logged. They note not just what you watch but when—if you watch cricket highlights at 11 PM on Thursdays, similar content will surface at the same time next week. Netflix can identify your mood within ten seconds of scrolling.

The systems don't need to understand why you binged three crime dramas in a row. They only need to see the pattern: repetition equals interest. Collaborative filtering groups you with millions of other users who exhibited similar behaviors, then predicts what you'll want next based on what they watched. The algorithm doesn't guess. It calculates.

Gaming platforms like Steam take this further, recording playtime, achievements, genre preferences, even controller input patterns. The "Because You Played" recommendation rows feel eerily personal because they are—but they're built on behavioral data, not demographic assumptions. The platform doesn't know you're a 34-year-old teacher. It knows you play story-driven indie games for 90 minutes on weekday evenings and prefer turn-based combat.

Who Benefits, Who Doesn't

The French study revealed an uncomfortable truth: while streaming increases overall diversity of consumption, it worsens cultural inequality. The platforms disproportionately benefit highly educated, high-income consumers who adopted the services first and have the digital literacy to navigate them effectively.

This creates a clear chain: digital inequalities lead to adoption gaps, which lead to cultural inequalities. Monthly subscription costs matter. So do the skills required to parse recommendation interfaces, manage multiple services, and actively curate watchlists. The same technology that expands horizons for some leaves others behind.

Being a "cultural omnivore"—someone with varied tastes across genres and forms—has been a status marker since the 1990s. Breadth of cultural consumption builds social capital. Streaming platforms accelerate this dynamic, giving already-advantaged consumers tools to diversify their tastes while others remain limited by cost, access, or comfort with the technology.

The Cost of Convenience

Critics worry that algorithmic recommendations encourage passive, superficial engagement. Users stop actively seeking out new music or films and instead let the platform surface content. The effort required to discover truly novel material—reading reviews, following artists, exploring outside your comfort zone—diminishes when a perfectly adequate suggestion appears automatically.

This matters beyond individual experience. Algorithmic personalization may erode shared cultural touchstones, the common references that help people connect. When everyone gets different recommendations, communal experience fragments. Your Spotify Discover Weekly playlist and mine have nothing in common.

Yet the data suggests people actually encounter more variety through streaming, not less. The paradox resolves when you realize algorithms don't just reflect existing preferences—they lower information costs. Recommendations help users navigate toward unfamiliar content they might never have found otherwise. The platform doesn't trap you in your tastes. It maps the territory just beyond them.

Beyond the Algorithm

Platforms don't just host content anymore—they actively sort, rank, surface, and hide it. This "algorithm culture" determines what becomes popular and what remains obscure, but not through some inscrutable black box. The mechanism is straightforward: behavioral data converted into structured variables like completion rate, engagement depth, and return frequency.

The real question isn't whether algorithms narrow or broaden our horizons. They demonstrably do both, depending on who you are and which medium you're consuming. The question is whether we're comfortable with cultural discovery being mediated by systems optimized for engagement rather than aesthetic value, novelty, or challenge.