Streaming engagement analytics

The short version

Netflix publishes what people watch, in two separate reports that don't line up. I joined them to answer a simple question: what's actually driving the hours?

The answer: people are watching more, but only because there's more to watch. The average title gets fewer hours than it used to.

Where the data comes from

Twice a year, Netflix lists every title and the hours it was watched. That covers about 99% of all viewing, but with no countries and no week-by-week detail.

Every week, it also publishes a Top 10 for the world and for 94 countries. That has the weeks and the countries, but only ten titles at a time. And for countries it's only a ranking, with no hours.

How it works

1. Keep the two reports separate. Squashing them into one table would mean inventing numbers that one of them doesn't have.
2. Make the titles match. The same show is written three different ways across the files, so those had to be lined up first.
3. Check it against Netflix. Before analysing anything, my totals had to match the ones Netflix published: 95.19, 96.21 and 97.66 billion hours.

What it found

• More titles, not bigger hits. Total hours went up each half-year, while the average title fell from 5.88 million hours to 5.62 million.
• Series beat films, and the gap is growing. A series earns about three and a half times what a film does. Films grew in number and still lost hours.
• Non-English films chart more and get watched less. More of them reach the global Top 10 than English films, and they earn less than half the hours.
• A new season brings the old one back. When The Gentlemen returned, its first season, off the global chart since May 2024, came back into the Top 10 of 56 countries that week and 85 the week after.
• Countries hold on to hits for very different lengths of time. A title stays in Ukraine's Top 10 for 5.8 weeks on average, and in Trinidad and Tobago's for 2.2.

What the data doesn't cover

Netflix doesn't publish hours by country, or genres. So anything about countries here is based on rankings, and it can tell a series from a film but not a comedy from a drama.

The dashboard is on this site, and the same tables load straight into Power BI.