1. Instead of saying "no", saying "not that way, but let's figure this out together".
2. Evaluating risk and modeling threats: "this is who we're protecting ourselves from, and here's what happens if we fail." If a bored teenager on their couch hacked our website, it would be embarrassing because someone without a lot of resources would be able to make changes to our display system, even if no real harm was done. If North Korea hacked our user database, it would suck and be bad for our users, but in practice not too many people are going to get angry at us for being attacked by a hostile nation's government as long as we were doing the right things.
(Note: that's grossly simplified, and it's not like we're "heh we don't protect against nation states".)
Anyway, I agree with your last sentence; at what point is something "good enough". Lately I feel like the "good enough" in a significant amount of corporations isn't acceptable. I'm in healthcare and the absolute lack of security in my day to day is absolutely amazing.
I agree with you on that last bit. While it's important to have your compliance ducks in a row, a lot of shops seem to feel like "we've checked all the audit checkboxes so we're secure now!" No. All that stuff is nice, but having a documented process for deciding who gets root on your database servers is not the same as actually securing your database servers.