How to Plot a Novel: Cause, Effect, and Every Payoff You Owe
Three drafts into my second novel, I finally admitted that chapters fourteen through seventeen were dead. Not badly written -- I could read a paragraph aloud and it sounded fine. Dead in the way a plot goes dead: my detective interviewed a witness, then interviewed another witness, then found a clue, then found a second clue, and none of it made the next thing happen. Each scene sat next to the one before it like items on a grocery list. I could delete any one of the four chapters and the story would not notice.
What fixed it was not a better witness or a sharper clue. It was a single sentence from a lecture E.M. Forster gave at Cambridge in the 1920s, later collected in Aspects of the Novel. "The king died and then the queen died" is a story, Forster said. "The king died, and then the queen died of grief" is a plot. The difference is one clause, and it turned out to be the entire difference between the four chapters I had and the four chapters I needed.
I have come to believe that plotting -- as distinct from outlining, which is really a scheduling problem -- is the discipline of making every event in your story a consequence of the one before it, and a cause of the one after it. Outlining tells you what happens in what order. Plotting tells you why that order is the only order that could have happened. Most craft advice quietly conflates the two, which is how a writer ends up with a beautifully organized folder of scenes that still reads like a grocery list. If you have not yet settled on a method for building that order in the first place, our guide to outlining a novel covers six of them in depth; this piece assumes you already have an order and is about the wiring underneath it.
Forster's Distinction Between a Story and a Plot
Forster's example is almost insultingly simple, which is exactly why it works as a diagnostic. "The king died and then the queen died" gives you two events and no relationship between them -- a timeline. "The king died, and then the queen died of grief" gives you the same two events, but now the second is caused by the first, and a reader does something a timeline never asks them to do: they start guessing. Why did she grieve so deeply? What kind of marriage was this? What happens to the kingdom now that neither of them survived it? One causal clause turns two facts into a machine that generates questions, and a plot is nothing more than a chain of those machines, each one producing the question the next scene answers.
You can test your own draft against this with a blunt instrument. Read your scene list and, between every pair of scenes, try inserting "and then," "therefore," or "but." If "and then" is the only word that fits, you have a sequence. If "therefore" or "but" fits, you have a chain. Trey Parker and Matt Stone, who have written and produced South Park since 1997, have described running close to this exact test on their own outline cards, replacing every "and then" between beats with "therefore" or "but" before an episode is allowed into production. A writers' room built around a weekly deadline does not keep a rule that fails to save it time, and this one survives because it catches the same dead scenes I found sitting in chapters fourteen through seventeen.
Escalation Means Consequences, Not More Obstacles
Kurt Vonnegut used to draw the shapes of stories on a chalkboard -- an axis running from Ill Fortune to Good Fortune, a line rising and falling across it as the plot moved through time. His favorite shape, "man in hole," is deceptively simple: a character falls into trouble, and the trouble gets worse before it gets better. What the drawing hides is the part that actually matters. The hole has to get deeper in a way that is different from how it got deeper last time, or the reader's sense of the line goes flat even while your incident count keeps climbing.
This was the exact mistake sitting in my four chapters. Each scene introduced a new obstacle -- a reluctant witness, a locked door, a missing file -- and every obstacle cost my detective the same thing: time. Time is a real stake. It does not compound. Losing three hours in chapter fourteen and three more in chapter fifteen does not feel worse the second time; it feels identical, which is why the chapters were interchangeable. Real escalation changes the currency. The reluctant witness should not just delay the detective -- she should make the detective doubt a colleague, which changes what the detective is willing to risk in chapter sixteen, which changes who she is willing to lie to in chapter seventeen. Each obstacle has to spend something the previous obstacle did not touch -- trust, money, a relationship, a piece of the character's self-image -- or the escalation is arithmetic instead of compound.
The steepest version of this problem usually shows up at the midpoint, the scene where a story's stakes either compound for the second half or quietly reset. The piece I wrote about what a midpoint reversal actually has to do is worth reading alongside this one -- it is the single moment where escalation either takes or fails, and readers feel the difference within a page even when they could not tell you why.
The Setup and Payoff Ledger Every Novel Keeps
Every setup in a novel is a debt. The gun on the wall in chapter two, the character who is "just a family friend," the odd scar the protagonist never explains -- each one is a promise that something is coming due. Chekhov's advice, often paraphrased from his letters, is close to the founding law of the form: if a rifle is hanging on the wall in chapter one, it has to go off by chapter three, or it should not be hanging there. The corollary matters just as much and gets ignored twice as often. Nothing should go off that was not, at some earlier point, hanging on a wall. A payoff with no setup is not a twist. It is a coincidence wearing a twist's clothes, and readers can feel the difference even when they cannot name it.
Ronald Knox, an English priest and one of the founding members of London's Detection Club, published a set of ten rules for fair-play detective fiction in 1929 that reads, nearly a century later, like a plotting manual disguised as a parlor game. The rule that has stuck with me longest is the one insisting that the reader be given every clue the detective has, at the moment the detective has it. Knox was writing about mysteries specifically, but the underlying demand -- that a payoff be traceable, in hindsight, to information the reader actually received -- applies to every genre. Your reader does not need to predict the ending. They need to be able to look back after it and find the setup already sitting there in plain sight, the second time through.
A quick way to check your own ledger: go through your manuscript and list, on one page, every setup you have planted -- objects, lines of dialogue, small unresolved questions -- next to the chapter where you plant it. Then, in a second column, note the chapter where each one pays off. Two kinds of gaps will show up immediately.
- Setups with no payoff. Anything you introduced and forgot -- the sister mentioned once in chapter three who never appears again, the locked drawer nobody opens. Either pay it off or cut it. A reader's attention is not infinite, and every unpaid setup spends some of it for nothing.
- Payoffs with no setup. The convenient confession, the document that surfaces exactly when the plot needs it, the skill the protagonist suddenly has in chapter twenty-eight. Go back and plant the setup earlier, even if it is a single sentence. A payoff seeded three chapters before it lands reads as inevitable. The same payoff with no seed reads as luck, and luck is the one thing a climax cannot survive on.
Coincidence Can Open a Plot, Never Close One
Robert McKee's rule in Story, his notoriously blunt 1997 screenwriting manual, draws a line between two kinds of coincidence that beginning plotters mix up constantly. Coincidence is allowed to start a story. A stranger sits down next to your protagonist on a train, a wrong number connects two people who should never have met, and the plot is under no obligation to explain why the universe arranged it. What coincidence cannot do, in McKee's rule and in my own experience revising three novels' worth of endings, is resolve a story. If your climax depends on the villain confessing for no clear reason, or the exact right document turning up in the exact right drawer at the exact right hour, you have spent your reader's trust at the one moment the plot most needed to be earned rather than arranged. Everything after the inciting incident has to be caused by choices your characters already made, not luck the author is quietly supplying from offstage.
This is also where subplots tend to go wrong. A romance thread, a rivalry, a secondary mystery -- each one makes its own promise to the reader, and each promise has to be paid off with the same discipline as the main plot, not resolved in a single convenient paragraph because the word count is running out. If a subplot's resolution depends on a coincidence the main plot would never be allowed to use, cut the subplot back until its ending is earned by the same causal chain as everything else in the book.
A Worked Example: Fixing a Subplot That Coincidences Its Way to an Ending
A few years ago I workshopped a manuscript with a subplot that illustrates the problem cleanly, because the writer had built her main plot with real discipline and let this one thread get away from her. The protagonist had an estranged brother, introduced in chapter two as a single passing mention -- he had not spoken to the family in six years, for reasons the book never specified. He reappeared in chapter twenty-nine, on the exact day the protagonist needed a car, a place to hide, and someone willing to lie to the police on her behalf, and he supplied all three without hesitation or explanation. The workshop's verdict was unanimous, and correct: nobody believed a word of it.
The fix was not to cut the brother. It was to move his causal weight earlier. We went back to chapter two and gave the estrangement a specific, concrete cause -- he had once covered for their father in a way that cost him everything, and swore he would never do that for anyone again. That single change did two things at once. It explained why he had been gone for six years, which the original draft never bothered to justify, and it meant that when he showed up in chapter twenty-nine and chose to help anyway, the choice cost him something a reader could feel, because twenty-seven chapters had already established exactly what helping had cost him the first time. The event on the page did not change at all. The causal chain underneath it changed, and that chain is the only thing that turns a coincidence into a payoff.
Notice what the fix actually required: not a new scene, not a twist, not additional research. One paragraph in chapter two, rewritten to plant a reason instead of a fact. That is usually the size of the fix once you have found the gap. The expensive part was never the rewriting. It is noticing, scene by scene, which gaps exist in the first place, which is exactly what the audit below is for.
A Practical Audit for a Draft That Already Exists
The writer who taught me to actually do this work systematically was not a novelist. John McPhee, who has spent six decades writing structural nonfiction for The New Yorker, has described drawing his pieces out as diagrams before writing a word of them -- not outlines in the school sense, but literal shapes on paper, with arrows connecting one piece of material to the next, each arrow representing a reason the next piece belongs exactly there and nowhere else. McPhee is writing about geology and oranges and the Alaskan wilderness, not fiction, but the diagram answers the identical question a novelist has to answer for every scene: what does this piece of material make possible that nothing before it made possible?
You can run a smaller version of the same exercise on a manuscript you have already drafted. Open your scene list -- not your outline, the actual scenes as they exist on the page -- and for each one, answer three questions:
- What does this scene know, or have, or believe, that the previous scene did not give it?
- What does the next scene need from this one in order to happen at all?
- If you deleted this scene, what specific later scene would stop making sense?
Scenes that fail question three -- the ones where nothing downstream would break -- are your chapters fourteen through seventeen. They may be well written. They may even be interesting in isolation. They are not plot. Give them a job, a piece of information, a decision, a relationship shift that a later scene genuinely depends on, or cut them and let their single best image survive inside a scene that already has one.
What Causation Costs You, and Why It Is Worth It
None of this makes drafting easier in the moment. Writing "and then" is faster than writing "therefore," because "and then" only requires you to know what happens next, while "therefore" requires you to know why it had to. Pantsers in particular sometimes resist this kind of audit because it feels like it is asking them to plan a novel they would rather discover. It is not. You can discover the causation exactly as freely as you discover the events -- the difference is that once you notice a scene is dead weight, on your second or third pass, you go back and either wire it into the chain or take it out, instead of leaving it in because it took a week to write and the sentences are good.
I eventually cut chapters fourteen and sixteen entirely and gave chapter fifteen's best image -- a witness who kept refolding a napkin into smaller and smaller squares while she lied -- to a scene in chapter nineteen, where my detective already had a reason to notice hands that would not stay still. The book that resulted is six thousand words shorter than the draft I abandoned, and it moves in a way that draft never moved, because every chapter that survived the cut is now something the next chapter needs. The queen still dies of grief, in every plot worth finishing. I had just spent three drafts figuring out who her grief was actually for.
Map your causal chain in Plotiar. A flowchart makes setups and payoffs visible as connected nodes instead of buried across a manuscript you have to reread cover to cover, and the plot grid tracks which thread owes which payoff before you reach the page where it comes due. Try it free.