Knowing when a draft is done is one of the harder practical questions in writing, and one of the least discussed. There is a great deal of advice about how to start writing and a reasonable amount about how to revise, but very little about how to stop — how to recognise the moment when further work is no longer improving the piece and the draft should be called finished.
This matters more than it might seem. A draft revised past its best is not the same as a draft that was never revised at all, but it has its own set of problems: the over-explained argument, the sentence that was once natural and is now laboured, the piece that has been smoothed into something careful but flat. Knowing when to stop is part of the craft.
The Signals I Look For
The clearest signal that a draft is done is when revisions stop changing the piece and start trading one version of it for another. Early in revision, a change improves something — a sentence becomes clearer, a section holds together better, a transition that was awkward becomes smooth. These changes are directional. The piece gets better.
Later in revision, changes stop being directional. A word is replaced with another word that is neither better nor worse — just different. A sentence is restructured and reads slightly differently, but not more clearly. A paragraph is moved and the piece is not improved by the move, only altered. When revision becomes this kind of lateral trading, the draft is usually done. Further work is not making the piece better; it is making it different in ways that have no clear rationale.
A second signal is when I can read the piece from beginning to end without stopping. Early drafts create friction — a sentence that needs rewriting, a section where the argument drops out, a transition that doesn’t hold. I feel these as small resistances when I read. When I can move through the piece without resistance, when nothing pulls me out of the reading, the draft is usually at the point where further revision is unlikely to improve it significantly. These are the two clearest signals I have that when a draft is done, it is genuinely done.
The Traps That Keep Writers Revising
The most common trap is revising to delay rather than to improve. A draft that is finished but not yet sent or published is still potential — it could still be brilliant, still be everything the writer wants it to be. The moment it is released, it is fixed, and its qualities become apparent and final. Revision is a way of keeping the piece in the state of potential, and writers who are anxious about the work tend to find reasons to keep revising even when the work does not need it.
The tell is a particular feeling: the sense that the piece is almost done, that there is something not quite right, but without any clear idea of what that something is. When revision is driven by a specific, identifiable problem — this transition is weak, this section is too long, this word is wrong — it is purposeful revision. When it is driven by a vague unease without a clear object, it is usually anxiety, not editorial judgement — and it will not tell you when a draft is done.
The second trap is revising for readers who have not yet read the piece. A writer imagines an objection, adds a qualification to meet it; imagines another objection, adds another qualification; imagines a reader who might not know a particular thing, and adds an explanation. Each addition seems justified in isolation. Together, they produce a piece that is over-qualified, over-explained, and stripped of the confidence that made it worth reading in the first place. The piece has been revised not toward what it needs to be, but away from the criticism it might receive.
The Discipline of Knowing When a Draft Is Done
The practical discipline I use is to make one final pass — a slow, careful read — and commit to it. Whatever changes I make in that pass, I make. Whatever I leave unchanged in that pass, I leave. The pass ends, and the draft is done.
This requires accepting that the piece will not be perfect. No piece of writing is, and a writer who waits for perfection does not finish. The question is not whether the piece could be improved — it always could be — but whether further improvement is available to this writer, at this moment, working on this particular piece.
When the answer is no, the draft is done. Not because it has reached some objective standard of completion, but because the writer has taken it as far as they can take it, and the next version would be different rather than better.
That is the line I try to hold. Not perfection, not a draft without problems, but knowing when a draft is done means taking it as far as it can go and is ready to be read. Calling it done is not giving up on the piece. It is completing it.
The piece exists now. That is not nothing.
