How to revise a novel

The draft exists. It is too long, the middle sags, and there is a subplot that ends twice. You open chapter one to begin fixing it, and you spend an afternoon on the first three paragraphs.

That afternoon is the problem. Not because the paragraphs did not improve, but because chapter one is throat-clearing that this revision is going to cut, and you have just spent your best hours of the week making the cut prettier.

Revision is three jobs. They want doing in order, and the order is the opposite of the one that feels natural.

Wait until you have forgotten it

Before any of it: put the draft down long enough that you stop remembering writing it.

Two to six weeks is where most writers land. Stephen King’s figure in On Writing is at least six, with the draft in a drawer and unread, so that it comes back feeling like the work of someone else who writes suspiciously like you. The point is not rest, it is amnesia. While you can still recall what you meant by a scene, you read your intention instead of your sentences. Gaps close themselves. A motivation you thought about for an hour and put on the page in four words reads as fully explained, because you have the hour in your head and the reader has the four words.

You cannot see what is missing from a page you can recite. Distance is the whole mechanism.

When you come back, read the entire thing in as few sittings as you can manage, and do not fix anything. Take notes — a line per problem, no solutions — because the moment you start rewriting you stop reading like a reader, and the read-through is the only time you will ever see the book the way someone encountering it does.

Pass one: the story

This pass is allowed to break things. It cuts chapters, merges characters, moves the ending, throws out the subplot. Nothing here is about sentences.

The questions are structural:

Does the story start where it starts? Most first drafts open a chapter or three before the story does, because the author needed a run-up. The reader does not. The test is where the situation becomes unstable — where something changes that cannot change back — and everything before that is usually setup you can distribute later.

Does the middle have its own engine? A sagging middle is almost always a middle where the protagonist has stopped choosing. Events happen to them, they respond, and the book idles. Look for the last decision they actively made and how far it is from the next one.

Does anything happen twice? Two scenes making the same point is the commonest structural fault in a draft, and it is invisible while writing forward because they were written months apart. One of them is better. The other one is gone.

Is the ending set up? Not foreshadowed — set up. Whatever resolves the book has to exist in it before it resolves anything, or the ending arrives as an announcement.

Cutting is the hard part, so use the mechanical test rather than judgement. Take the scene out. Read the scene before and the scene after, back to back. If nothing is missing, it stays out. Usually what is missing is one piece of information, and it fits in a sentence somewhere else.

King’s rule of thumb for how much comes out is the second draft as the first draft minus ten percent. Treat it as an expectation rather than a target: it tells you that a revision which removes almost nothing has probably not been a structural pass at all.

Keep the cuts. A file of deleted scenes makes cutting psychologically cheap, and cheap cutting is most of what a structural pass needs. You will restore about one in twenty and it does not matter, because the cost of keeping them was only disk space.

Pass two: the scenes

Now the shape is settled. This pass asks whether each surviving scene does a job.

A scene that works ends somewhere different from where it started. Someone knows something they did not know, wants something they did not want, or can no longer do something they could do at the top. If a scene ends in the same position it opened, it is an event rather than a scene, and events belong in summary.

Two related checks are worth running mechanically across the whole book:

Enter late, leave early. Most drafted scenes begin with arrival, coats, greetings and a little weather before anything starts, and continue past the moment they resolve. Look at your first paragraph and your last: the real scene is often three paragraphs in and ends a paragraph before you stopped.

Whose scene is it? Every scene belongs to somebody who wants something in it. When a scene feels flat and you cannot say who that is, that is the diagnosis.

This is also the pass for pace, because pace is a property of scene lengths in sequence rather than of prose speed. Long, long, long, long is a slog no sentence-level editing rescues. The fix is structural: shorten one, cut one, break one in half. Chapter length is the same instrument at a larger scale.

Pass three: the line

Only now do sentences matter, and only now is it safe to spend an afternoon on three paragraphs, because those paragraphs are staying.

What a line pass is actually looking for, in rough order of how much it improves a book:

Filtering. She saw the door open is weaker than the door opened, because the first puts a camera between the reader and the event. Draft prose is full of she felt, he noticed, she realised, and cutting them is the single change that most reliably makes prose feel closer.

Adverbs propping up weak verbs. Walked quickly against hurried. The adverb is not the problem; the verb is.

Sentences that all run the same length. Read a page aloud and you will hear it. Rhythm is the part of prose that dies in revision, because each sentence gets fixed in isolation and they all converge on the same shape.

Dialogue that answers the question asked. Real speech evades, changes the subject, and answers a different question. Drafted dialogue tends to be characters efficiently exchanging information.

Repeated words within a few lines of each other. Your ear catches these when reading aloud and never when reading silently.

Reading aloud is the whole technique here, and it is worth doing for the entire book even though it takes days. Your mouth finds what your eye skips.

What proofreading is not

Proofreading is not a fourth pass of revision. It is a check for errors — typos, a character’s eyes changing colour, a Tuesday that follows a Thursday — and it happens after the text is final, ideally by someone who is not you, on a version that looks different from the one you drafted in.

That last part is practical. Reading the book in a different form breaks the pattern-matching that lets you skim your own sentences. Reading it as typeset pages rather than as a manuscript file is one of the more effective ways to do this, and it doubles as a check that the book works as a book: chapter openings, running heads, where the pages break.

Then, and only then, the formatting

The book stops changing at the end of revision, and that is the moment production can start.

The order matters because it is expensive to get wrong. Trim size decides how much text fits on a page. Typography decides it again. Together they produce a page count, and the page count decides the inside margin, the spine width, and what the printer charges per copy. Every one of those is downstream of a text that has stopped losing chapters.

A book revised down from 95,000 words to 82,000 is a different physical object: fewer pages, a narrower spine, a different cover file, a lower unit cost. Doing the interior before the text settled means doing it again.

In Nueel that resettling is cheap — change the trim or the type and the whole book repaginates, so you see the new page count before you commit to anything — but the cheapest version is still the one you do once, on a finished text. Word count to pages covers what that arithmetic looks like before you have a file.

---

Most stalled revisions are not stalled on difficulty. They are stalled because the writer is doing pass three on a book that has not had pass one, and no amount of careful sentence work makes a structural problem go away. Fix what the book is, then what its scenes do, then how it reads. The afternoons you spend on paragraphs will be spent on paragraphs that survive.

Sources

• Stephen King, On Writing: A Memoir of the Craft (2000) — the six-week drawer, the closed-door first draft, and the second draft as the first minus ten percent

Common questions

How do you revise a novel?

In passes, from largest to smallest. First the structure: does the story work, is anything missing, does anything need to go. Then the scenes: does each one earn its place and end in a different position from where it started. Then the line: the actual sentences. Doing these in the other order means editing prose that later gets cut, which is the most common way revision consumes months without improving a book.

How long should you wait before revising a first draft?

Long enough that you read it as a reader rather than as the person who remembers writing it. Two to six weeks is the usual range, and Stephen King's advice in On Writing is at least six with the draft unread. The purpose is not rest, it is forgetting: while you can still recall what you meant, you read your intention rather than what is on the page, and you cannot see gaps that your memory keeps filling in.

How many drafts does a novel need?

There is no correct number, and counting drafts measures the wrong thing. What matters is whether the passes are distinct. Three passes with different jobs will do more than eight where you fiddled with sentences each time. Most novels need one structural rebuild, one or two scene-level passes, and a line pass, plus a proofread that is not revision at all.

What is the difference between revising and editing?

Revision changes what the book is: scenes are cut, added, reordered, characters are merged, a subplot goes. Editing improves how it is written without changing what happens. Revision is the expensive one and has to come first, because line-editing a scene that revision will delete is work you pay for twice.

How do you know what to cut?

The reliable test is whether the story still works without it. Cut the scene, read the two scenes on either side back to back, and see whether anything is actually missing. Most scenes that feel unnecessary are, and what feels lost is usually information that can move into a neighbouring scene in a paragraph.

Should I revise as I write?

Not the same material, no. Rereading yesterday's page to get back into the voice is fine and most writers do it. Rewriting chapter three while chapters four to thirty do not exist is different, because you are polishing something whose job may change entirely once the book has an ending. Nothing about a scene is final until you know what it leads to.

When should I format the book?

After revision, when the story has stopped changing. Trim size, margins and typography all depend on a stable text, and the page count that follows from them decides the spine width and the printing cost. Formatting a book that is still losing chapters means doing the interior twice.