← All guides

Revision, demystified

How many drafts does a novel need?

The short answer: there's no magic number — and authors who ask "how many drafts?" are usually asking the wrong question. What matters isn't the count, it's whether each pass had a job and finished it. Most novels get publication-ready through three to five deliberate passes. Here's what each one is for.

The four passes that matter

PassThe jobIgnore for now
1. Discovery draftGet the whole story down, start to finishEverything else. Ugly is fine; finished is the only goal
2. Structural passFix the story: where it starts, the sagging middle, arcs, subplots, the endingSentences — no point polishing scenes that may not survive
3. Scene & line passMake every scene earn its place, then tighten the prose: filter words, echoes, rhythmTypos — they're caught later
4. Polish passConsistency, continuity, a clean spelling-and-grammar sweepSecond-guessing settled structural decisions

Some books need the structural pass twice. Some authors split the line work into two rounds. That's normal — the pattern flexes. What doesn't flex is the order: story before scenes, scenes before sentences, sentences before polish. (It's the same logic that governs what order professional edits go in.)

Why counting drafts misleads you

"Draft two" for one author means a page-one rewrite; for another it means fixing typos. The number tells you nothing — the completed job does. A more useful question after each pass: what class of problem did I just eliminate? If you can't answer, the pass was churn: rereading, tinkering, swapping commas — motion that feels like work but moves nothing.

How to know when you're done

  • Your changes are swaps, not fixes. Rewording a line you reworded last month — and half-preferring the old version — means solo revision has hit diminishing returns.
  • You're reversing earlier decisions. Un-cutting scenes, re-cutting them. The book isn't improving; it's oscillating.
  • The problems you can name are all small. When nothing structural is bothering you, either it's fixed — or you can no longer see it. Both mean the same thing: time for outside eyes.

Outside eyes, in cost order: an objective diagnostic first (the free Scorecard below), then beta readers on the revised draft, then professional editing on the strongest version — cheapest feedback first, most expensive last. Before any of it, run the self-editing checklist so nobody's time is spent on problems you could have caught yourself.

Find out where your manuscript actually stands — free.

The Manuscript Scorecard grades your full manuscript on nine craft metrics, benchmarks it against professionally-edited books, and tells you exactly what to fix first. Free, in minutes.

Get your free Manuscript Scorecard →

FAQ

Is one draft ever enough?

For publication, effectively no. Even very experienced authors who draft clean still run structural and polish passes — they've just internalized enough craft that their passes go faster, not that they skip them.

Can a book have too many drafts?

Yes — the failure mode is oscillation, not improvement: reversing earlier changes, endlessly rewording. When your edits stop fixing and start swapping, more solo drafts won't help; outside feedback will.

Should each draft be a complete rewrite?

No. Except for a discovery draft with a broken premise, a 'draft' is a targeted pass with one job — structure, scenes, lines, polish. Full rewrites are occasionally necessary, but they're the exception.

When should an editor enter the draft cycle?

After your structural pass at the earliest — an edit of a draft you already know is structurally broken pays professional rates for old news. Diagnose cheaply first, revise, and spend the professional money on your strongest draft.

More guides