The Revision Strategy Writing Guide: Turning a First Draft into a Finished Novel

The most paralyzing belief a writer can carry is that a good first draft is possible. It is not — not because writing is uniquely hard, but because first drafts are acts of discovery, and discovery means going down paths that turn out to be wrong, writing scenes that reveal the story doesn't work the way you thought, and building characters who turn out to be different from the characters you planned. A first draft that flows perfectly from beginning to end is a first draft where the writer was not discovering anything. Anne Lamott called it the shitty first draft. Hemingway said the first draft of anything is garbage. These are not self-deprecating jokes — they are craft statements. The first draft's job is not to be good. Its job is to exist, so that revision has something to work with. A finished bad draft is infinitely more valuable than an unfinished good attempt. This means the goal in drafting is completion, not quality. Write through the scenes you are not sure about. Write placeholder text and move on. Write the version of the chapter where you know something is missing but cannot figure out what yet. The missing thing will become apparent in revision, when you can see the whole. What is impossible to see from inside the draft — the structural problems, the pacing issues, the characters who are underdeveloped — becomes visible when you can look at the whole manuscript at once. Giving yourself permission to write a broken first draft is not lowering your standards. It is understanding what drafting is for. The standards come in revision, and revision cannot happen until the draft exists.

The most common revision mistake is trying to fix everything at once. The writer who opens their first draft intending to simultaneously fix the structure, deepen the characters, improve the pacing, and tighten the prose will make slow progress on all fronts and decisive progress on none. Revision is most effective when it is stratified — when you move through the manuscript in distinct passes, each focused on a specific level of the work. The layered approach works from the largest scale to the smallest. Start with structure: does the story work as a whole? Before you spend time perfecting a scene in chapter four, confirm that chapter four needs to exist. Before you polish dialogue, confirm that the scene the dialogue is in is doing necessary work. Large-scale cuts and restructuring render small-scale fixes moot. Sequence matters. Once structure is sound, move to scene level. Does each scene work on its own terms: clear goal, conflict, change, and consequence? Is the pacing within the scene right? Does each scene do what it needs to do for character and plot? Only then move to the line level: prose quality, word choice, rhythm, clarity. There is a practical reason for this order beyond efficiency. If you polish the prose before confirming the structure works, you become emotionally attached to beautifully written sentences that need to be cut. The polish becomes an argument against the surgery. Do the surgery first, when you are still objective about what needs to go. Polish what survives. Give yourself time between passes. Distance from the manuscript is a prerequisite for clear eyes.

Structural revision is the most important and most feared pass. It requires seeing the whole manuscript as an object — a machine with parts, some of which are functioning and some of which are not — and making decisions about what needs to be moved, removed, or rebuilt. This is the pass where chapters get cut, POV characters get dropped, and the order of events gets rearranged. It is also the pass where most writers resist the most, because the scale of the changes required can feel like failure rather than revision. To conduct a structural revision, start with a scene-by-scene outline of what you actually wrote — not what you planned to write, but what exists on the page. List every scene: what happens, whose POV it is, what changes by the end. Then read the outline with fresh eyes, as if it is a story you are considering whether to continue. Where does it drag? Where does the logic break? Where does a scene exist only because you needed it for backstory, not because it earns its place? Look specifically for three structural problems. First: a weak midpoint. The midpoint should be a genuine turning point that splits the second act into two distinct movements. If the middle of your novel is just “more of the same,” the story will sag. Second: an unearned climax. Has the story set up everything the climax requires? Is the reader adequately invested in the outcome? Third: a resolution that doesn't match the promise. The ending should answer the question the story has been asking since the first chapter. If it answers a different question, something in either the opening or the ending needs to change.

Once the structure is confirmed, move into each scene individually and ask a suite of diagnostic questions. These questions apply regardless of genre or subject matter: they are about whether the scene is functioning as a narrative unit. First: what does the point-of-view character want in this scene, and is that goal clear? A scene without a clear character goal is a scene that has no spine. The reader will drift without knowing what they are watching for. Second: what is the scene's conflict? Not necessarily an argument or a fight — but some form of resistance between what the character wants and what the scene provides. If the character wants something and gets it with no friction, the scene is too easy. If the character wants something and is prevented from getting it by nothing more than plot convenience, the scene is inert. Third: does the scene change anything? At the end of the scene, something should be different from how it was at the beginning — a relationship has shifted, a piece of information has been revealed, a decision has been made, a character's understanding of their situation has changed. If the scene ends exactly as it began, it is a dead scene and should be cut or reconceived. Fourth: where does the scene start and end? Most scenes start too early, with characters arriving at the location and establishing themselves. Cut to the moment of conflict. Most scenes end too late, with characters processing what just happened and winding down. End at the point of change. The trimmed version is almost always more effective.

Line-level revision is the most granular pass, and it comes last for good reason: lines serve scenes, and scenes serve structure. A perfectly written sentence in a scene that does not need to exist is wasted craft. Only when you are confident the scene is working do you invest in making each sentence in it as good as it can be. Line-level revision has several distinct concerns. Clarity: does each sentence mean exactly what you intend? Read sentences in isolation and ask whether their meaning is unambiguous. Rhythm: read the prose aloud. Does it flow, or does it catch in places? Varied sentence length creates rhythm; monotonous sentence length creates drone. Economy: every word should be doing something. Adverbs that modify verbs that are already specific (“walked quickly” instead of “strode”), adjectives that state the obvious, throat-clearing constructions at the start of sentences (“There was a”) — cut them. Specifically, look for the passive voice where the active would be stronger, and abstract nouns where concrete ones would be more precise. “She had a feeling of sadness” versus “she sat on the edge of the bed for a long time without moving”: the first names, the second shows. Also look for filter words — “she saw,” “he heard,” “she felt” — that put a layer of distance between the reader and the experience. In close third or first person, most filter words can be cut without losing meaning, and the prose becomes more immediate. Line revision is also where you listen for your own verbal tics — the words and constructions you return to habitually. Every writer has them.

Beta readers are one of the most valuable revision tools available, and one of the most misused. The most common misuse: sending the draft too early, before structural revision, in the hope that reader feedback will tell you what is wrong. This approach wastes the beta reader's time and gives you information you already know — that the draft is broken — without the specific structural clarity you need to fix it. Send to beta readers after structural revision, when you are confident the story works as a whole and you are seeking feedback on how it lands emotionally and narratively for someone who is not you. At this stage, beta readers can tell you things you genuinely cannot know from inside the manuscript: where they lost interest, which characters they cared about and which they didn't, where they were confused, what they wanted more of. Choose beta readers carefully. They should be readers of your genre who can articulate their responses, not just friends who will be kind. The most useful feedback is specific and emotional: “I stopped caring about the protagonist in chapter six and I don't know why” is actionable. “It was really good overall” is not. Ask specific questions rather than “what did you think?” — the open question produces the least useful responses. Ask about pacing, about character investment, about the moment they were most engaged and the moment they most wanted to put it down. Read the feedback with patience: wait a few days before responding, let defensiveness subside, and then look for the patterns in what multiple readers report independently.