Streaming ratings are a fiction

The streamer viewership numbers that appear in press releases and trade coverage are, with rare exceptions, marketing artifacts rather than reliable data. Netflix's Top 10, Amazon's viewing statistics, Apple's occasional disclosures, and the internal metrics the streamers quote to justify renewals are calculated using methodologies that have shifted repeatedly, are audited by nobody outside the platforms themselves, and are chosen specifically to produce numbers that make the platform look good.

This is not a conspiracy theory. It is a straightforward observation about the difference between the Nielsen system that governed broadcast television — imperfect, but externally audited and comparable across networks — and the streaming era's balkanised, self-reported, methodologically inconsistent viewership disclosure. The trade press has largely stopped pointing this out because the platforms are the sources for most of the coverage, and criticising the sources produces friction. But the data problem is real, and the industry has organised itself around numbers that are not, strictly speaking, numbers.

The consequences are largest in commissioning. If a streamer's internal metric of choice is a two-minute-view threshold, then the shows that get commissioned will be optimised for two-minute-view retention rather than for anything else. If the metric is completion percentage, then shows will be optimised for completability. Every metric produces a specific set of incentives, and the streamer that chooses the metric shapes the commissioning slate in ways that no external actor can audit or contest.

The counter case is that the streamers have been more transparent recently than they used to be, and that the trade press does now push back on the more absurd numbers. Both are true. But the transparency is still marginal, and the pushback is still infrequent, and the underlying problem — that the industry lacks a shared measurement standard — has not been resolved. Netflix's Top 10 is not comparable to Nielsen's ratings, and it is not comparable to anything Amazon or Apple release. The absence of comparability is a structural weakness.

What the industry needs is what it had for most of the twentieth century — an independent, audited, cross-platform viewership standard, funded by a coalition of studios and streamers and operated by an entity with no commercial stake in the results. Nielsen has tried to move into the streaming era and has been rebuffed by the platforms. The Movielabs consortium has proposed alternatives. None of them has traction, because the platforms benefit from opacity. That benefit is the reason the reform has not happened, and the reason it should.