[willchang]

There's a good chance Netflix is behaving rationally.

Shows are like startups. The payoff is a very skewed distribution, i.e. median and mean are vastly different. After one season, I'd argue you have a pretty good estimate of the earning potential for a show. Maybe you're even off by a factor of two or three, but since the goal is to find a show with 10x or 100x the viewership, you're better off canceling a show with median success.

Actually it's probably a lot easier to predict the payoff for a show after one year, than for a startup. Startups can take multiple years to find product-market fit; shows rarely pivot dramatically.

[Update: I got the facts wrong with the following example, as someone nicely pointed out below! I think the general principle still stands.] Tuca & Bertie, for example, was canceled after one season, but the same people went on to make Bojack Horseman, which was a much bigger hit. Bojack would not have happened if Tuca were renewed, and Tuca would never have gotten as popular as Bojack.

[rideontime]

You've got your timeline reversed, Bojack predated Tuca & Bertie by years.

[CameronNemo]

Plus Tuca & Bertie is the best show Netflix has ever produced.