The number. Among new shows that launched a first season in 2024 on Netflix, Disney+, Max and Prime Video, seasons of three to six episodes recorded the highest average completion rate: 48 percent. Seasons of eleven to fifteen episodes recorded an average of 26 percent. Sixteen episodes and above rises to 29 percent.

The source. Digital i, “Are You Still Watching?”, a content trend report published in May 2025 from its viewing panel across the territories it measures; the public summary carries the figures, the full report is gated. Averages per title, first seasons only, 2024 launches. Digital i describes the figures as an average per title of the share of completers, does not define completion on the public page, and does not state what the share is taken over, so they are reported here as it labels them, a completion rate, and nothing is claimed about how many people left. It attributes the rise at sixteen episodes and above to those seasons often being soaps or anime. The public page does not report the band of seven to ten episodes, where most streaming first seasons sit. A search of Nielsen, Antenna, Ampere, Omdia, Luminate, Parrot Analytics, Samba TV, Netflix’s own engagement reports, the trade press and the academic literature found no other published completion rate by episode count, and no trade outlet has reported these figures independently, so this is one measurement from one panel. The same report counts 279 original shows launched in 2024 against 395 in 2022.

What it says. A long first season asks more of the audience before it gives the platform anything back. On these four streamers in 2024, a show with eleven to fifteen episodes was finished, on average, at about half the rate of a show with three to six. That matters because the finale is where a subscription earns its next month: a viewer who stops at episode seven never reaches the episode that was meant to keep them. The small rise above sixteen episodes is the clue to what is going on. Digital i says those long seasons are often soaps and anime. [inference] Their audiences chose the length on purpose, so the length cost them nothing. Length hurts when the audience did not sign up for it.

Where the data ends. Four things this number cannot do. It is an average across many titles, not a curve for any one show, and inside a single length band the spread is wide: in Digital i’s 2022 figures for Netflix in Europe, as reported by What’s on Netflix, every title ran eight to ten episodes and completion ran from 41 percent to 87. It cannot separate length from the things that travel with it: genre, episode running time, whether the season dropped at once or weekly, and how long the panel waited before counting. No published source pulls those apart. It says nothing about India, because no Indian platform publishes completion by episode count in any language. And it cannot tell you whether a particular long season should have been short, or whether the story needed the length and the audience was the wrong one for it. That call is made before episode one is written, with no completion curve to look at.

What I would decide anyway. The measurement is completion by episode, and every platform already has it for its own titles, whether or not it publishes it. Before the next episode count is signed off, pull that curve for your own last ten first seasons and put it next to the count the writer is asking for. The decision it changes is the order, not the story. If your own long seasons lose the room before the finale, commission the shorter order and buy the second season once the first has earned it. If they hold the room to the end, this finding does not apply to you, and your own completion rate on those ten seasons is the number that says so.