I think it's just a reality that in the era of streaming very few people are going to stick out a mediocre series in the hopes that it has enough good bones that the showrunners will eventually do something with them.
Netflix made a lot of promises about what data would enable. There's very little indication it's helped Netflix to develop better shows.
I'm arguing that isn't always true. Yes, some shows knock it out of the park in season 1. Other shows take a few seasons to really hit their stride.
It appears Netflix is taking a SV/VC approach to shows. Spend big, fail fast, and count on a small number of massive hits to succeed. While this may pad revenue in the short term, I'm not convinced it's the best long-term play. If subscribers get annoyed at short production runs and inconsistent programming, they leave and go to Disney or Hulu or Amazon.
Of course, I'm not in the industry at all, so I could be wrong. But, as a consumer, it is annoying to have Netflix push a few shows really hard, only to have them cancelled after 8 episodes.
It's easier to stick when you are early in the game and the marketplace is less crowded. Hose of Cards got a lot of extra publicity for being the first.
If they're good enough it will stick. See stranger things, house of cards, and orange is the new black for examples. The problem is that Netflix isn't very good at picking winners. They pay way more than anyone else for content and the results aren't there.