[gav]

It's a hard problem to solve: getting to some useful amount of "done" where it solves actual problems for customers is a long way from delivering a truly polished, quality product.

There's diminishing returns from 80% to 100% done, plus there's the human condition--it's a lot more fun to solve the hard problems than do the polishing at the end. I think that teams can get too enamored of the features they create, it's hard to get perspective when you've been staring at something for a while.

I've often wondered if the answer is to have multiple teams (pods) so that one team can get a feature to MVP, then hand it over. You're then not invested in "your" product, there's space to be objective and revisit the path from MVP onward.