Your manuscript is a mess. Chapter three contradicts chapter nine, one character's name is spelled two different ways, and there's a scene you skip every time you open the file. You know something has to change. The harder question is whether that something is the writing or the fixing. Drafting and editing use different parts of your attention, and they undermine each other when you run them at the same time. Here is how to tell which one you need right now.
Two jobs that don't share a brain
Drafting has one job: make something exist. Editing has a different job: make that thing work. The first rewards momentum and a certain cheerful carelessness. The second rewards patience, distance, and a willingness to be unkind to your own sentences.
When you edit while drafting, you tend to protect the wrong things. You polish a paragraph that will be cut in the next draft. You make a weak scene read smoothly enough that you stop noticing it's weak. Worse, you use up the decision-making energy you needed for the scene you haven't written yet. Most stalled drafts aren't stalled because the writer doesn't know what's wrong. They're stalled because the writer is trying to diagnose and build at the same time, like framing a wall while the foundation is still wet.
So the first decision isn't what to fix. It's which mode you're in for the next hour.
Signs you should keep writing
Keep drafting when any of these are true. They don't all have to be true; one is usually enough.
- You don't know how it ends. Even with a detailed outline, the ending on the page is rarely the ending in the plan. You can't judge structure you haven't finished building.
- You're well short of your target length. If you're at forty percent of a first draft and already fixing sentences, you're decorating a house with no roof.
- Your notes are outpacing your pages. When new ideas keep arriving faster than you can write them down, that's a drafting signal, not a repair signal. Follow it.
- The fix would cascade. If solving a problem means changing things you haven't written yet, you're guessing. Note it and move on.
- You've rewritten the same paragraph five times. That's not editing, it's fidgeting. Put a placeholder in and keep going.
- You keep rereading the opening chapters. Rereading is often a way of avoiding the scene that frightens you. Write the frightening scene badly instead.
Placeholders are your best tool here. Use a bracket and a plain instruction: [argument here — she should say the thing about her father]. Write the note as if you were briefing a colleague. A draft with fifteen honest placeholders is more useful than a draft with a beautifully edited first act and nothing after it.
Signs it's time to start fixing
The balance tips when the problems stop being local and start being repetitive. A few reliable indicators:
- You can summarise the whole thing in three sentences without lying about what's on the page.
- The same note appears in the margin four or five times. One recurring note is a pattern, and patterns are structural.
- New pages feel like padding. You're writing around a hole rather than into one.
- You dread opening the file. Boredom during revision is data; if a scene bores you on the tenth read, it will bore a reader on the first.
- You've reached an ending, even a bad one. A finished bad draft is a diagnosable object. An unfinished good draft is not.
Notice that none of these say "the prose is clumsy." Clumsy prose is the last thing you should worry about, and often it's clumsy because the scene beneath it is doing two jobs or none.
What to do when you genuinely can't tell
There's a middle stretch where the draft is too far along to write freely and too unfinished to revise properly. It feels like wading. Two techniques help.
Write an outline of what's actually on the page
Not the outline you planned — the one you wrote, scene by scene, in one sentence each. Keep it to a page. Then read the page instead of the manuscript. Structural problems that are invisible in three hundred pages become obvious in thirty lines: the subplot vanishes for a hundred pages, two scenes do the same work, the midpoint changes nothing.
Make a problem list, not a fix list
Name the problems in plain language and stop there. "The antagonist has no reason to wait until the last act." "The middle sags because she has nothing to lose until chapter twenty." A problem stated clearly usually suggests its own solution within a day or two. A solution invented on the spot while you're still drafting usually creates a new problem you'll have to unpick later.
The order of operations
When you do start editing, the sequence matters more than the effort. Each pass should look through a single lens, because switching lenses mid-pass is how you end up with a polished sentence in a scene that shouldn't exist.
- Finish the draft. Placeholders allowed, weak scenes allowed, no polishing.
- Let it rest. As long as you can bear, ideally a few weeks. You need to read it as a stranger, and strangers don't remember what they meant.
- Read once without changing anything. No pen, no track changes, just note how you feel at each chapter break. Boredom and confusion are findings.
- Structural pass. Scenes, order, point of view, what's at stake. Expect to cut and move large blocks. This is the pass that most changes the book.
- Scene pass. Does each scene turn? Someone wants something, something gets in the way, and the situation is different at the end than at the start. If a scene only delivers information, find a way to make it cost something.
- Line edit. Sentence by sentence: rhythm, repetition, clarity, the removal of throat-clearing. Only now does the prose get your attention.
- Proofread. Read aloud, or read the last sentence first and work backwards, so your brain stops predicting the words.
The most expensive mistake: editing at the wrong level
You can spend an entire weekend smoothing the dialogue in a scene you'll cut on Monday. That's the trap. Line editing a scene that's going to be deleted, restructured, or reassigned to a different character is work you'll throw away, and it makes the deletion harder because the prose now glows.
The fix is stubbornness about level. In the structural pass, refuse to touch sentences. Read them as symptoms, not as problems. If a paragraph is confusing, ask what the scene is trying to do there, not how to reorder the clauses. Save the clause-level care for a pass where nothing above it can move. It feels slow at first, and then it saves you weeks.
The test to use tonight
If you can't state the problem in one sentence, you're not ready to fix it. Write the sentence in the margin — plainly, no jargon — and keep drafting. The sentence will still be there tomorrow, and it will be more useful then than six paragraphs of pre-emptive repair.
Messy is a stage, not a diagnosis. Every finished book passed through a version that embarrassed its author. Your job is to know which tool is in your hand: the one that makes pages, or the one that makes them better. Use one at a time, and the mess becomes a to-do list.
Photo: Pixabay / Pexels
