How to plan scenes in a novel
· The StoryProp team
Ask a stalled novelist where the book went wrong and they will usually name a chapter. But a chapter is a typographic convenience — a place for the reader to put the book down. The working unit of a novel is the scene: a continuous stretch of action in which somebody wants something, meets resistance, and comes out changed. Planning at the scene level is what separates an outline you can actually draft from one that sits in a drawer while you improvise anyway. The most durable tool for the job is the scene and sequel method, and its practical companion is the scene list — one line per scene, the whole book on a few pages. Neither requires index cards or a weekend seminar. Both require you to answer one question about every scene before any other: what changes?
Plan the change before the content
Most scene planning fails because it records content instead of change. “Mara and Daniel argue at the kitchen table” is content: it says what the scene contains, not what it does. You could write that argument for four pages and end with everything as it was — two people annoyed with each other, which they already were. The scene happened; the story didn’t move.
A change-first plan reads differently. “Mara arrives certain Daniel has never set foot in the house; she leaves having seen the chipped cup on his shelf.” Now the scene has a job: certainty becomes doubt. Whatever dialogue you invent, whatever business with the kettle and the coats, the scene is finished when that reversal lands, and not before. A useful discipline is to state every planned scene as an arrow — trust → suspicion, safety → exposure, resolve → temptation. If you can’t write the arrow, you don’t yet have a scene. You have a location and some characters waiting for a reason to exist.
The scene and sequel method
The scene and sequel method — set out by Dwight Swain in Techniques of the Selling Writer — formalizes the arrow into two alternating units. A scene, in this narrow technical sense, has three beats. Goal: the point-of-view character wants something concrete, now — not “happiness” but “get Daniel to admit he kept a key.” Conflict: something resists — another person, a locked drawer, the character’s own cowardice — and the resistance escalates rather than repeats. Disaster: the scene ends worse, or at least otherwise, than the character intended. Not always catastrophe; the technical term covers any turn that closes off the easy path. Yes, he kept a key — and so did her mother.
The sequel is the unit nobody plans and every book needs. After a disaster, the character reacts — pure feeling, however briefly rendered. Then comes the dilemma: thinking through options that are all bad, because if one option were obviously good the disaster wasn’t trying hard enough. Then a decision, which is to say a new goal. Reaction, dilemma, decision. The sequel is where readers digest what happened and where characters earn their next move; skip it and the novel reads like a highlight reel — things keep happening to people who never metabolize any of it.
The alternation is the engine. The decision that closes a sequel becomes the goal that opens the next scene; the disaster that closes a scene provokes the reaction that opens the next sequel. Cause and effect are built into the structure itself, which is why books planned this way rarely drift into “and then… and then… and then.”
Sequels are elastic. A thriller can compress reaction, dilemma, and decision into one paragraph of white-knuckled arithmetic; a quieter novel can give a single sequel a whole chapter of walking and remembering. That ratio of scene to sequel is one of the main dials controlling how fast a book feels — a diagnostic subject of its own, taken up in how to fix pacing in a novel. For planning purposes the rule is simpler: the sequel must exist. It can be three sentences. It cannot be skipped after any disaster the reader actually felt.
How to write a scene list
A scene list is one line per scene, in order, for the whole novel — the closest thing prose has to an x-ray. The flesh of the writing is stripped away and what remains is the skeleton of change. If the bones are wrong, no amount of sentence-level polish will make the body stand up; and unlike the manuscript, the skeleton fits on a few pages, where your eye can travel the whole book in a minute.
The format matters less than the fields. A workable line carries the point-of-view character, the goal in one clause, the opposition in one clause, the disaster or turn, and the arrow. Give the sequel its own short line, ending in the decision it produces. Something like: “Mara / get the locksmith’s records / the clerk stonewalls, then lets slip a name / the name is Daniel’s / trust → suspicion. Sequel: she decides to say nothing to her sister.” That is enough to plan from and enough to diagnose with.
Read the list in columns, not rows. Scan the disasters first: if three consecutive scenes end with “they agree to keep looking,” you have found the stretch where the book idles. Scan the goals: “find out more” and “deal with the situation” are placeholders for decisions you haven’t made yet, and they will stall the draft exactly where they sit. Scan the arrows: if one value flips back and forth — trust → suspicion → trust → suspicion — the plot is pacing its cell, not traveling.
Keep the list current whether you plot in advance or discover as you go. Built beforehand, it prescribes; built chapter by chapter behind the draft, it diagnoses. Either way it makes cutting cheap. Deleting a scene from a manuscript costs an afternoon and leaves scars; deleting a line from a list costs a second, and the orphaned decisions on either side show you instantly what the cut would break.
Enter late, leave early
Once every scene has its arrow, plan its edges. The old cutting-room advice applies to prose: arrive as late as you can, leave as early as you can. If the scene’s job is the collision between Mara’s question and Daniel’s evasion, then the drive over, the parking, the doorbell, and the coats are all pre-scene. Open with the question already asked — or with the cup already noticed and neither of them mentioning it.
Leaving early is harder, because endings feel as if they want de-escalation — goodbyes, a door closing, a settling paragraph. Mostly the scene should end within a breath of its turn: the disaster lands, cut. White space does the settling, and the sequel handles the emotion deliberately instead of letting it leak away in pleasantries.
Plan the edges in the list and you stay honest at the desk. A note as small as “in: mid-argument / out: on the cup” will save you three warm-up pages you would only cut later.
From list to draft
The scene list’s last job is shaping the drafting itself. Sitting down to write chapter twelve is an invitation to dread; sitting down to write “Mara / the attic / find her mother’s letters / the floor gives way / hidden → exposed” is a task with edges. The change is decided, the entry and exit are marked, and the invention left to do is the good kind — texture, voice, surprise inside a container.
Tools can respect this unit or fight it. AI novel-writing software built for long projects will draft in scene-sized pieces at your direction rather than pouring out chapters wholesale; StoryProp, for instance, drafts piecemeal by default and reads your recorded decisions — who knows what, which arrows have already fired — before it writes, so the scene you planned is the scene that arrives. What you agree to gets written down; what you don’t stays a suggestion.
No one holds a novel whole in their head; it is too long, and the writing of it takes too many months. The scene list is how you see the book whole, and the arrow is how you enter it small. Plan the change before the content, walk in late, get out on the turn, give the sequel room to breathe — then move down the list to the next line, where somebody wants something and is about to be told no.
A scene list only pays off if the drafting keeps faith with it. StoryProp writes one scene at a time from the decisions you have recorded, so the arrow you planned is the arrow that lands on the page.