The problem is not so much whether shows are killed to early or too late, but when you blindly follow metrics to determine either way. There are plenty shows p

The problem is not so much whether shows are killed to early or too late, but when you blindly follow metrics to determine either way. There are plenty shows pre-streaming that got killed prematurely (Firefly) and plenty that dragged on way past their prime (Simpsons) because of exec idiocy or internal politics or whatever that had little to do with whether the quality of the shows was good or whether they had found an audience or just needed time to grow. I think Netflix tried to avoid that by using data-driven analysis to determine whether to keep or kill a show, but it turns out you probably need some human in the decision chain with an experienced gut. Another problem is that you don't necessarily know whether a show is going to be popular upfront, so the writers try to have a self-contained story arc within a season. If it gets renewed you end up having to awkwardly start a new arc with characters that were pretty much "done" in terms of development. This is pretty obvious with Jessica Jones and Stranger Things, both of which were arguably much weaker in their second seasons.