On one of the LinkedIn Groups, Steve Green, who I value a lot both for his experience in exploratory testing as well as a person argued against the use of exit criteria as they are almost meaningless in his experience.
I agree but meaningless doesn't mean useless. Let's say that:
The discussion and learning about "it" is more important than the definition of "it".
While in many cases exit criteria are sacrificed to get a project out of the door I think that it would be a mistake to dismiss them. By trying to write down exit criteria and discussion within the team a lot can be learned about the way software is developed within a given company. When one struggles to come up with an exit criteria that may point to one or several problems.
Are stakeholders not clear? Are decisions made ad hoc? Is planning or estimating a problem?
A discussion about what exit criteria may be realistic and which ones aren't including the reasoning behind it will unearth a lot of important information for the project.
You may end up without exit criteria after all and that's OK. But the important bit is that information was gained and learning happened.
Of course "it" could be anything. Exit Criteria; a bug; a process; a workflow; an approach; a standard. As long as there's learning we have gained something that cannot be taken away.