The Streaming Wars Were Won on Content. The Next Ones Won't Be.

Streaming’s original deal with the consumer was beautifully simple: everything, anywhere, for less than cable. One app, one subscription, one search bar. It solved a genuine problem — access — and for a while it was brilliant. No more broadcast schedules. No more paying for 200 channels to watch five.

Then every studio looked at that success and drew the obvious, wrong conclusion: we should do that too. Each launched its own service. The content spread across platforms, the aggregate price went up, and the experience went down. The average streaming household now runs somewhere around eight services and $1,100 a year — for a discovery experience that is, by most measures that matter, worse than the cable bundle it replaced.

Here’s the version of this I live, most weeks. Kids in bed by 8:30, roughly ninety minutes of evening left. We spend twenty of those scrolling through four different apps. By nine o’clock we’ve watched nothing. And the thing is — that’s not a content problem. There has never been more great content. It’s a friction problem.

Recommendation isn’t context

The diagnosis I keep landing on: every platform does recommendation. None does context.

Consider two moments in the same person’s week. Wednesday, 1pm: phone, office, twenty-five minutes. Friday, 8pm: living-room TV, surround sound, the evening ahead. Ask the apps what to watch in each. Wednesday’s answer is everything — an undifferentiated wall of thumbnails. Friday’s answer, in my house, should obviously be the Saturday-adjacent family programme we watch together every week. What I actually get, in both moments, is the same grid — plus a promoted tile for a programme I have never watched a single episode of and never will. That’s not personalisation. That’s editorial promotion wearing personalisation’s badge.

Context-aware means the platform knows the difference between a lunch break and a Friday night: device, time, viewing history, who’s likely in the room. The technology to make that decision per viewer, per moment, inside the delivery path is exactly the edge-inference architecture the AI infrastructure world is busy building — which is why I keep insisting media is one of its first serious customers.

Piracy is a demand signal

When finding what you want legitimately is this hard, some people go somewhere it’s easy. The industry mostly frames that as an enforcement problem, and enforcement matters — content authentication, forensic watermarking, provenance credentials are table stakes and getting better. But enforcement alone is playing defence.

The sharper framing: piracy is a demand signal, not a moral failing. It tells you exactly where your experience is broken. When the pirated experience is faster and simpler than the paid one, the fix isn’t only in the courts — it’s in the product. Reduce the friction, add the things piracy can’t offer (offline, social, live), and make paying the easy path. Protect and solve, simultaneously; neither works alone.

The experience stack

The framework I use to sort investment in this space is a four-layer stack, split into two tiers.

Table stakes — necessary, undifferentiated, and invisible when done well:

• Deliver. Efficient global delivery at scale — multi-CDN, per-title encoding, edge caching. The infrastructure constant.
• Protect. Enforcement technology that makes piracy harder and traceable — watermarking, content credentials, authentication.

Where you win — the layers where margin and loyalty actually live:

• Personalise. Context-aware delivery that knows the difference between a lunch break and a Friday night — edge compute, manifest control, context signals.
• Monetise. If you can personalise the experience, you can personalise the revenue — server-side ad insertion, edge ad selection, privacy-preserving targeting.

The uncomfortable audit question for any streaming platform: where does your investment actually sit? If it’s overwhelmingly in Deliver and Protect, you’re optimising table stakes — spending to not-lose rather than to win. The content layer changes every quarter; the experience stack is the constant. (If you want to feel the trade-offs directly, the Experience Budget Allocator lets you split a latency budget across origin, regional, edge, and client tiers and watch what each millisecond costs.)

The bottom line

The content wars had a clear scoreboard: catalogue size, exclusives, spend. That war is over — not because someone won it, but because everyone spent themselves into the same exhausted position and the consumer stopped rewarding the next billion dollars of library.

The next war is scored differently. Time-to-play. Discovery abandonment. Browse-to-play conversion. The minutes my family doesn’t spend scrolling. These are measurable, instrumentable, fixable numbers — and almost nobody is competing on them yet.

The streaming industry solved access. Now it has to solve experience. The platforms that survive won’t be whoever spent the most on content. They’ll be whoever made it easiest to find and watch what you actually want.

The full framework — including the decision checklist for platform teams — is in the From Content Wars to Experience Wars insight.