Well I guess maybe my wording wasn't ideal. Sometimes it's 'what they want changes significantly through the development process due to outside factors'. Having time to iterate is good. Being able to experiment and figure out what's valuable is good. Being given a strict deadline, then halfway through being expected to chuck out nearly everything because the client brought the demo to the rest of the board and the latter didn't like it/chewed them out for it is perhaps not so great.
Yes, but I'd argue that that bad thing is downstream of the implicit expectation people seem to have that people do and especially that they should know what they want, up front. In this example, the client thought they knew what the board wanted, but they didn't, and you thought you knew what the client wanted, but you couldn't possibly, because what they wanted was a thing the board would like, but they clearly didn't know what that was.
It's a sticky wicket because of course time in front of the board is precious so you don't want to constantly run things by them and make them micro-manage, but I think the board would have probably chewed them out less if they had brought a tiny MVP (or even just a proof of concept) that hadn't required significant investment, and asked "here's what we have with almost no investment, here's our plan for the next small step, what do you think of this direction?".
