The Revision Triage: How to Find the Problems That Matter Most in a Novel

A first draft can contain perfectly good sentences and still fail as a novel. It may have vivid characters, an appealing setting and several memorable scenes, yet feel slow, confusing or emotionally thin when read from beginning to end. This is why revision becomes frustrating when it starts with wording. You can improve a paragraph that will later be cut, polish dialogue in a scene that belongs somewhere else, or spend an afternoon repairing a chapter whose real problem began eighty pages earlier. A better approach is revision triage: identify the highest-level problem first, then work down towards scenes, paragraphs and sentences.

A better approach is revision triage: identify the highest-level problem first, then work down towards scenes, paragraphs and sentences. The aim is not to make the manuscript flawless in one pass. It is to stop spending careful attention on work that may not survive the next decision.

Begin with a reading that does not invite editing

Before you revise, read the manuscript in a form that makes you a reader rather than a mechanic. Print it, export it to an e-reader, or change the typeface and page size on screen. The alteration matters because a familiar document encourages you to remember what you intended instead of noticing what is actually there.

Do not correct as you go. Keep a separate document or a sheet of paper for observations. Note where you became curious, where you felt lost, where your attention weakened, and where an emotional moment did not have the effect you expected. Page numbers are useful, but brief descriptions are more valuable than a long list of typos.

Use questions such as:

What does the viewpoint character want at this point?
What prevents them from getting it?
What changes by the end of the scene or chapter?
Why does this event need to happen here?
What information is the reader waiting for?
Which parts did I skim, and what caused me to skim them?

Resist the temptation to defend the manuscript. A confusing chapter may be confusing for a good reason: you have withheld information to create suspense. It may also be confusing because the causal link between two events is missing. At this stage, record the effect before deciding whether the effect is intentional.

Sort problems into three levels

Most revision notes become manageable when they are sorted into three levels: story, scene and prose. These levels interact, but they should not be tackled in the same order.

Story problems affect the novel as a whole. They include a weak central conflict, an ending that does not answer the book's main question, a character who has no meaningful choice to make, or a subplot that never affects the main action. If the central relationship, investigation or struggle lacks direction, sentence-level improvements will not solve the difficulty.

Scene problems affect individual units of action. A scene may repeat information, begin too early, lack a clear pressure point, or end without changing the reader's understanding. A chapter can be beautifully written and still be unnecessary if the same work is done more effectively elsewhere.

Prose problems include repetition, vague description, clumsy syntax, weak verbs, inconsistent tense and dialogue that sounds artificial. These matter, but they are usually the last layer. Editing them first often creates a polished version of material you later have to restructure.

Revise in the order of consequence: fix the decision that changes the most pages before fixing the sentence that changes one.

Find the manuscript's governing question

Many novels become difficult to revise because the writer has several descriptions of what the book is about, and none of them helps with decisions. A governing question is more practical than a marketing summary. It identifies the unresolved pressure that keeps the reader moving.

It might be: Can a disgraced surgeon accept responsibility before the person she harmed exposes her? Or: Will two estranged brothers protect their family if doing so requires them to reveal the secret that divided them? The question should contain an active conflict and imply that the answer will be earned through events, not supplied by explanation.

Write your own question in one sentence. Then inspect each major section of the novel and ask whether it complicates, advances or answers that question. A section may contribute through a secondary relationship or a change in circumstance, but it should have a discernible connection. If a sequence cannot be connected without a paragraph of explanation, it may be carrying over from an earlier version of the book.

This exercise is not an argument for making every chapter narrowly functional. Fiction needs texture, contradiction and breathing space. The test is not whether every scene advances the external plot. It is whether the scene alters something that matters: a relationship, a belief, a risk, a plan, a piece of knowledge or the reader's understanding of a character.

Make a scene inventory before making cuts

Once you have identified the book's governing question, create a simple inventory of its scenes. You do not need a sophisticated software system. A table with one row per scene is enough. Record:

The viewpoint character.
The location and approximate time.
What the character wants at the start.
The obstacle or opposing pressure.
The important action or discovery.
What is different at the end.
How the scene connects to the next one.

Keep the descriptions short. “Mara argues with her brother about the missing money” is more useful here than a detailed synopsis. The inventory should show the shape of the manuscript at a glance.

Look for rows that have no clear desire, obstacle or change. These are candidates for cutting, combining or rebuilding. Also look for runs of scenes that perform the same function. Three conversations in which a character learns that the investigation is dangerous may be individually competent but collectively repetitive. The strongest version may combine the best information with the highest personal cost.

Pay special attention to scenes that end with a revelation but do not produce a decision. Information only becomes dramatic when somebody must do something with it. If a character discovers that a rival has lied, what do they choose next? If the choice appears in the following chapter, consider whether the discovery and decision belong in the same scene.

Distinguish missing scenes from weak scenes

A common revision mistake is to keep rewriting a scene that cannot work because the manuscript has skipped a necessary step. The character may seem to change their mind without an experience that would persuade them. A relationship may become intimate between chapters. A dangerous plan may appear fully formed, even though the reader has not seen the argument or discovery that made it possible.

When a scene feels abrupt, ask whether it is poorly written or whether the book is missing an earlier event. Signs of a missing scene include:

A character knows something the reader has not seen them learn.
A decision has emotional weight but no visible cause.
A relationship changes mainly through summary.
A location, object or threat becomes important without introduction.
The next scene assumes a consequence that the previous scene did not create.

The missing material does not always need to become a full chapter. A brief exchange, a physical action or a carefully placed observation may be enough. The point is to supply the cause, not to add length for its own sake.

Revise character arcs as a chain of choices

Character development is often described in abstract terms: confidence, grief, trust, freedom. Those words are useful themes, but they are difficult revision tools. A character arc becomes easier to examine when you track choices.

List the major decisions made by your viewpoint character. For each one, note what they believe, what they fear and what they are trying to protect. Then compare the early choices with the late ones. Has the character's behaviour changed because circumstances changed, or because their understanding changed? Are they making harder choices, or merely receiving more severe consequences?

An arc does not require a character to become nicer, braver or more agreeable. A tragic arc may show a person recognising the truth and refusing to act on it. A comic arc may involve someone surrendering a rigid belief while keeping their sharpness or ambition. What matters is that the ending grows from the pattern of choices established earlier.

If the final decision feels unearned, do not immediately add an explanatory speech. Look for earlier moments in which the character could have acted differently. Give them opportunities to practise the value, weakness or compromise that the ending will test. Readers are more likely to accept a transformation when they have seen its pressure building in behaviour.

Only now work at paragraph and sentence level

After the structural pass, read again with a narrower purpose. You can now edit prose with greater confidence because the scenes are less likely to vanish.

Start with clarity. Replace vague references, untangle pronouns and check that the physical action can be followed. Then look at emphasis: does each paragraph know what it is doing? A paragraph may establish a setting, reveal a thought, create a threat or change the rhythm of an exchange. If it tries to do all four at once, its main effect may be blurred.

Next, examine repetition. Writers often repeat an idea in several forms because they are trying to make sure the reader understands it. Trust the strongest expression. Look for repeated explanations of motivation, repeated descriptions of a character's appearance, and dialogue that states what both speakers already know.

Finally, read aloud. The purpose is not to perform the prose theatrically. It is to hear where sentences stall, where dialogue becomes too formal, and where every paragraph has the same length and cadence. Mark the places that resist speech, then decide whether the problem is syntax, punctuation or simply an overloaded idea.

Use a revision plan with a stopping point

Revision expands to fill every available week unless you define what “done” means for the current pass. Choose one pass and give it a limited brief. For example:

Pass one: clarify the governing question and remove scenes that do not support it.
Pass two: repair missing causes, weak choices and repeated scene functions.
Pass three: strengthen viewpoint, description and dialogue.
Pass four: check continuity, grammar and presentation.

Keep a “later” list for attractive distractions. Perhaps a chapter would benefit from a different image, or a minor character could have a more interesting occupation. Record the idea without following it immediately. This protects the current pass from becoming another first draft.

At the end of each pass, read the relevant section again from the beginning. Do not judge the whole novel after every change. Look for evidence that the specific problem has improved. If it has not, return to diagnosis rather than applying more polish.

Revision is a sequence of decisions

The most useful question in revision is not “How can I make this better?” It is “What kind of problem am I looking at?” A story problem may require a new decision or a changed ending. A scene problem may require a cut, combination or missing cause. A prose problem may require a sharper verb or a cleaner sentence.

Separating those levels makes revision less mysterious. It also makes the work more economical. You stop asking language to rescue structure, stop adding scenes to compensate for unclear motivation, and stop mistaking familiarity with the manuscript for understanding it.

A novel rarely improves through one grand correction. It improves when the writer makes the right decision at the right level, then gives that decision time to reshape the pages around it.