There are two kinds of done
Shipping something that works and actually finishing it are not the same thing, and the gap between them has a price you pay later.
Yesterday I spent most of the day going back.
Not because something broke. Nothing broke. Everything worked. But I'd built a set of documents for one of the businesses I run, and somewhere between "shipped" and "actually looked at them side by side," inconsistencies had piled up. One element didn't match the template. Another was close but not right. A third had a structure that almost fit the original, but not quite.
Individually, none of it was wrong. Collectively, it was sloppy. And sloppy compounds.
"Done is better than perfect" is true. I believe it. Shipping something imperfect is almost always better than waiting for flawless — because flawless usually never arrives, and while you wait, nothing runs. But there's a version of that quote people use as a permission slip for carelessness, and that version is a trap.
Two kinds of done
The first is done-enough-to-move: shipped, working, in the world, ready to iterate. That's the right done. Healthy done. The kind that keeps momentum alive.
The second is done-enough-to-avoid-thinking-about-it: the small inconsistency you didn't quite close, the assumption you left loose, the thing that technically functions but creates quiet friction every time someone encounters it. That's not done. That's deferred.
The distinction matters because the second kind doesn't stay deferred. It sits there accumulating context — you have to re-learn what you were doing, re-understand the problem, re-stitch together why something is the way it is. The hour you didn't spend closing it becomes two hours later. Or four. Or a customer noticing before you do.
I've learned to notice the difference by feel. There's a different texture to finishing something I actually finished versus finishing something I'm pretending is finished. The first has a kind of cleanness. The second has a low hum — a frequency you can't locate but know is there.
Yesterday was a day of clearing that hum. No new feature. No launch. Just closing what I'd left open.
Sometimes the most important work is not building forward. It's actually finishing the things you told yourself you already finished.