I think that even at the 90% mark, there's value in giving what I would call "craft-work feedback": the type of feedback one craftsman gives another.<p><i>First</i>, you tell them that, <i>from the perspective of the end-user</i>, their craftwork is <i>done</i>, and nobody will notice anything wrong with it. It's shippable. It's usable. They probably won't hear one complaint, and sales will be good.<p><i>Then</i> you point out the little things. The things that only other craftsmen will notice, but that end-users, despite not knowing to <i>look</i> for them, would still <i>appreciate</i> if you changed. The things that would take the thing from "done" to "perfect." The things that are, to most craftsmen, "just a matter of pride."<p>But you must <i>finish off</i> by reminding them that since they have a deadline (and they do, even if they call it "runway"), it's not even a matter of fixing <i>all</i> the little things. They don't have <i>time</i> to fix all the little things. Instead, it's a matter of picking the highest-impact ones they <i>can</i> fix in the time available, fixing those, and then shipping.<p>I think the key insight from such an explanation is that there are always some issues you point out that they must simply "accept": it's wrong, and now they're aware it's in there, but it's <i>not</i> going to get fixed, and it's going to ship like that, and everyone is going to see it (though only other craftsmen will notice it.) Tell them that they can't be paralysed by this thought: instead, they must accept it, learn from it, and do better with the next one.