Yeah it's hard to develop a general criteria. There are also situations where you may be making a lot of invisible progress, and what appears to be a hopeless situation for years hits a breakthrough and suddenly huge progress is made overnight because people were actually listening to some degree. So trying to map everything out ahead of time can be a fool's errand when you have no objective measures of progress.
If the goal is to avoid sunk cost fallacies, I'd say the answer is same as anywhere, budgeting. Develop some general, convenient (very important they're convenient, otherwise you won't track them) metrics, so that way you have at least a general idea of how much time, money, sleep, even vague things like energy level, stress and opportunity costs can be tracked to some degree. That way you have some quantifiable measure of how many resources you're sinking into a project and if it's worth it (ie I was abnormally stressed out directly due to this effort for 137 days last year, is it worth another 137 days of abnormal stress? Or from an opportunity cost perspective I could have spent those 50 hours playing with my kids, would that have been a better investment?)
Also set some hard budget ceilings and very stiff criteria (like at least an 80% chance of meaningful progress, whatever that means specifically for the situation) for going over them. That way you don't feel trapped by the ceiling, but it still holds you accountable.
Of course it's all on the individual to maintain the metrics and stick to the budget, but I'd argue if you don't have the discipline to do even that you might want to question your ability to execute any plans for more complex community development :)