In the days of broadcast television, everyone knew how to measure a hit: the Nielsen rating, an estimate of how many households tuned in. Streaming changed that. Each service now decides what to measure and what to reveal, and the result is a flood of numbers that look precise but are hard to compare.
Netflix: from hours to “views”
Netflix publishes a weekly Top 10 list of its most-watched films and shows. Since 2023, it has ranked them by “views,” which it calculates by taking the total hours people spent watching a title and dividing by the title’s running time. A two-hour film watched for a total of 20 million hours counts as 10 million views.
The change was made partly to compare titles of different lengths more fairly: under an hours-only system, a long series with many episodes had an advantage over a short film. Netflix also publishes a much larger engagement report twice a year, listing viewing hours for nearly everything on the service.
What “views” doesn’t tell you is how many different people watched, or how many finished. Someone who watches half a film counts as half a view.
Nielsen: share of all TV time
Nielsen, the company behind traditional TV ratings, measures viewing on television sets through panels of households and other data. Its monthly report, called The Gauge, shows what share of all TV viewing time goes to broadcast networks, cable channels and streaming services.
Streaming’s share has grown steadily, and in 2025 streaming passed broadcast and cable combined as a share of TV viewing time for the first time. YouTube, viewed on TV screens, regularly ranks as the largest single platform in the report.
Nielsen also publishes weekly streaming top 10 charts based on minutes watched on TV sets. These include shows across several services, which makes them one of the few independent ways to compare titles on different platforms. But they measure only viewing on televisions, not phones or laptops, and there is a delay of several weeks.
Why the numbers are hard to compare
- Different metrics: views, hours, minutes, households and “accounts that watched” measure different things.
- Different time windows: a show’s first three days, first week or first 28 days can produce very different totals, and companies choose the window that looks best.
- Different screens: some numbers include phones and computers; others count only TV sets.
- Selective disclosure: services publicize their hits and often say little about shows that underperform.
When a press release calls a show “our most-watched ever,” it is usually true within the specific measure and time window chosen, and nearly impossible to verify against competitors.
What actually decides a show’s future
Inside streaming companies, success is measured with internal data that is rarely shared: how many people finish a season, whether a show brings in new subscribers or keeps existing ones from canceling, and how much it cost to make. A show with modest public numbers can be renewed if it is cheap and loyal viewers finish it; a show with big opening numbers can be canceled if viewers drop off.
How to read a streaming headline
- Check which metric is being used, and whether it is views, hours or households.
- Check the time window, especially for comparisons with older shows.
- Look for an independent number, such as Nielsen, when a headline compares platforms.
- Remember that completion rates and cost, which matter most for renewals, are almost never public.
What we think
Streaming numbers have become part marketing and part measurement. Services are more open than they used to be, Netflix in particular, but each still chooses the figures that flatter its own shows.
The healthiest approach is to compare like with like: views against views over the same number of days on the same service. Comparisons across platforms are best treated as rough, and the true measure of a hit, whether people stick around, mostly stays behind closed doors.
