There's definitely ~10% that aren't worth welcoming back, but I'm playing interlocutor here.

There's plenty of companies who tossed all their cash cows into volcanoes for the difficult customer belief which really stemmed from an inability to listen. Here, read up on the fall of WordStar for a premiere: https://en.m.wikipedia.org/wiki/MicroPro_International

Old software companies were famous for this: Lotus, VisiCorp, Digital Research - running away from their existing customers who they thought were atrophying to some presumed greener pastures that didn't exist only to end up with nobody because the problem was they had their heads up their asses the whole way.

In this problem space there's 3 levels which are both of increasing accuracy but also increasing work.

1. Imagine your customers 2. Measure your customers 3. Know your customers and almost-customers, like personally.

3 doesn't scale but ignoring it is what can snag even the biggest companies