Many a UX researcher arrives at a growing company and finds it is their full time job to try and identify everything that has been broken along the way. By the time companies get round to hiring full-time researchers, that's one hell of a lot of debris.

There's an acceptable level of 'known issues' that can appear in a new release, but that's why engineers invented the concept of 'technical debt' - it does need to be paid eventually.

I'm guessing half the reason companies slow down is not because they're suddenly afraid to make mistakes, but because they've already made too many.