A writer hands us a manuscript and apologises before we have read a word. The opening chapter is rough, she says, she knows it, she nearly rewrote it four times before sending it. When we read the chapter, it is rough — clotted, hedged, circling its own point for three paragraphs before saying anything. But the second chapter, written fast and never fussed over, is clean. This happens often enough on the editing desk that it has stopped surprising us. The pages a writer agonised over are frequently worse than the ones she dashed off, because agonising and drafting are different jobs, and trying to do them at once does neither well.
The first draft is not a worse version of the final draft
There is a tempting but wrong way to picture writing: the finished piece exists somewhere, fully formed, and each draft is a clumsy attempt to copy it down, getting closer each time. On that model a bad first draft is simply a failure of transcription. You knew what you meant; you just botched it.
That is not what is happening. In most cases the writer does not yet know what she means. The first draft is where the thinking gets done — not recorded, done. John McPhee, who has written for The New Yorker since 1963 and is about as far from a blocked amateur as a writer can be, described his own first drafts bluntly in his New Yorker essay on the writing process: “Blurt out, heave out, babble out something — anything — as a first draft.” He calls the result “that first awful blurting,” and notes that the difficulty never goes away with experience. The blurt is not a defective draft. It is the only way he can find out what he has.
This reframes the badness entirely. A first draft is bad in the way wet clay is bad at being a bowl. It is not failing at the task; it is at an earlier stage of a different one.
Why it has to be bad: the desk is too small
The mechanical reason first drafts come out rough is that good writing asks the brain to do several incompatible things at once. Drawing on Alan Baddeley’s model of working memory, the psychologist Ronald Kellogg has spent decades mapping what writing demands of the mind, and his 2013 progress report in the Journal of Writing Research sums up the finding: planning ideas, translating them into sentences, and reviewing what you have written all draw on the same limited pool of executive attention. They compete. When you try to generate a sentence and judge it in the same breath, neither gets full resources, and both suffer.
A first draft that is allowed to be bad is simply one where you have stopped running all three processes simultaneously. You let translating happen and postpone reviewing. The output looks worse on the page precisely because you spent nothing policing it — which is the point. The editing capacity you would have burned suppressing clumsy sentences is instead available later, when there is something to edit.
This is why the overworked opening chapter so often arrives broken. The writer was planning, drafting, and judging every clause at once, and the bottleneck shows. The chapter written quickly reads better not because the writer tried less, but because she let the jobs take turns.
What editors actually see in a draft that was protected from judgement
In manuscripts we work on, the difference between an honest first draft and a self-censored one is visible from the first page. The protected draft is messy but alive: it overreaches, repeats itself, follows tangents that turn out to matter. There is something there to cut toward. The self-censored draft is the harder problem. It is tidy and dead. Every sentence is defensible and none of them risks anything, because the writer killed every interesting impulse before it could embarrass her. You cannot edit courage back into a paragraph that never had any.
Anne Lamott made the canonical case for letting the first attempt be terrible in Bird by Bird (1994), where she calls it the “shitty first draft” and lays out a sequence editors recognise instantly: “The first draft is the down draft — you just get it down. The second draft is the up draft — you fix it up.” The order is not arbitrary. You cannot fix up what you have not got down, and trying to do both on the first pass is how writers end up with three immaculate sentences and a deadline.
Drafting badly on schedule beats waiting to draft well
If the first draft is where thinking happens, the practical question is how to keep producing them rather than waiting for the good version to arrive intact. Here the research is useful but also more contested than productivity books admit, and it is worth being honest about that.
The famous evidence comes from the psychologist Robert Boice, whose 1980s intervention studies are still cited everywhere. In one, academics who wrote in their usual saved-up binges produced far fewer pages over a year than those who wrote in short daily sessions, and those who wrote daily and reported to someone produced more still. Boice also found that writers forced onto a schedule generated more usable ideas, not just more pages — drafting bred thinking, rather than the reverse.
But the cleanest account of those studies is also their sharpest critique. In a 2016 paper in the International Journal for Academic Development, the University of Auckland’s Helen Sword went back to Boice’s original numbers and found his cohorts tiny, self-selected, and never replicated. Her own survey of more than 1,300 academics found that roughly seven in eight successful writers do not write every day at all. Her conclusion is not that scheduling fails — she keeps a daily routine herself — but that the iron rule does not hold. What survives is narrower and more reliable: regular contact with the page produces drafts, and drafts are the raw material. Whether that contact is daily or weekly matters less than that it keeps happening before the writing feels ready.
The bad draft is an asset, not a confession
Not every rough draft becomes good. Some manuscripts are bad in the down-draft way — too much, too tangled, alive — and some are bad in the way that no amount of editing repairs. The point is that the first kind is the kind you want, and it only exists if the writer was willing to be bad on the page on purpose.
So when a writer apologises for a rough chapter, the apology is usually aimed at the wrong target. The roughness is not the problem. The problem, when there is one, is the chapter she rewrote four times until it was safe — the one where she did all her editing before there was anything to edit, and left us a clean surface with nothing underneath. The bad draft, the one she is embarrassed by, is the one we can work with. It is the only one that was honest about not knowing yet.
