How to create conflict in a story
· The StoryProp team
A familiar note from editors: the pages are clean, the people are believable, and nothing is happening. The usual repair is an argument — someone slams a door, someone raises a voice — and the scene gets louder without getting tighter. An argument is a symptom of conflict, not the thing itself. Conflict is two wants that cannot both be satisfied, under enough pressure that somebody has to act. Every clause of that sentence is load-bearing. Wants, plural: one character wanting something is a goal, and a goal alone produces errands. Cannot both be satisfied: the wants must be genuinely opposed — or believed to be, which works nearly as well and fails more interestingly. Under pressure: without a reason to act now, opposed wants coexist politely for decades. Families are full of standing oppositions that never become scenes, because nobody has to move.
Opposed wants under pressure
Consider a small domestic example. In “The Quiet House,” Mara wants to sell the family house — she needs the money, and she needs to stop being the person the house happened to. Daniel wants to keep it exactly as their mother left it, down to the chipped cup on the drainboard. Neither want is unreasonable. Neither is as negotiable as its owner pretends. And for years nothing came of the opposition, because nothing forced the question.
Then a buyer makes an offer that expires at the end of the month, and the estate's costs come due. The offer doesn't create the conflict — that existed the day their mother died. The offer applies pressure, and pressure converts a standing disagreement into a sequence of scenes. When a manuscript seems to lack conflict, the missing ingredient is usually not opposition but a clock: some reason the characters can't keep doing what people mostly do, which is wait each other out.
The three story conflict types — and why they braid
The old classroom list — character against character, against the world, against the self — survives in nearly every guide to creating conflict in a story because it maps something real: where the opposing want lives. Internal conflict is one person holding two wants that exclude each other; Mara wants the money and wants not to become the sister who sold the house. Interpersonal conflict sets want against want between people. External conflict sets a want against something that can't be argued with — an institution, an illness, a storm, a deadline. What the list doesn't say is that these are not a menu. They're strands, and a story running on a single strand thins out fast.
Purely external conflict becomes logistics: the flood rises, the characters cope, competence is displayed. Purely internal conflict becomes rumination, a character circling the same decision in progressively finer prose. Purely interpersonal conflict becomes a squabble with no floor under it. The strength is in the braid, where each strand raises the price of the others. The estate deadline — external — forces Mara and Daniel into the same rooms. The rooms force the interpersonal collision. The collision exposes the split inside Mara that she'd been managing by staying away. Pull any strand and the others tighten.
Braiding also answers where conflict comes from: it comes from character. A want with history behind it — formed by an old wound, covered by habit — will find something to oppose without your help. That construction is the territory of want, mask, and pressure in character work; build people whose wants are specific and funded by their past, and collisions follow.
Friction in every exchange
Conflict is fractal. The book has its central opposition; each scene should carry a smaller one; and at the smallest scale, almost any exchange of dialogue can hold friction — two people wanting even slightly different things from the same conversation. One wants reassurance; the other wants to leave. One wants to confess; the other wants not to have heard. Nobody raises a voice. Nothing gets thrown.
Watch it in miniature. Mara comes into the kitchen holding the buyer's paperwork. Daniel is at the sink, washing the chipped cup with unusual care. She asks whether he's had a chance to look at the offer. He tells her the cup was their mother's favorite, and that the chip happened during a move, the year Mara left for college. Every line of his is warm, informative, and a refusal. Every line of hers is patient, reasonable, and a demand. The scene reads as quiet. It is wall to wall with conflict, because neither person will enter the other's conversation.
The working test for any scene is two questions: what does each person want from this exchange, and who gets it? If everyone wants the same thing and receives it, the scene is a handoff of information wearing dialogue as a costume. Cut it, or give someone in the room a want that points the other way.
False conflict: loud scenes with nothing under them
Two counterfeits show up constantly. The first is bickering: opposition without stakes. The characters disagree because the scene needed noise, but either could concede without losing anything that matters, so the reader files the exchange as texture and reads faster. The second is the obstacle without cost: the missed train, the locked door, the storm — delay that consumes pages and changes no one. The character arrives late and identical.
The test for load-bearing conflict is cost and change. Cost: if this character loses, do they lose something hard to recover — money is the weakest version, standing is better, self-image is best? Change: does the outcome move the situation — a decision made, a secret exposed, a relationship reweighted? A quarrel that ends where it began was decoration. A quiet scene that ends with one character knowing something she can't unknow was conflict, whatever the volume.
A third counterfeit deserves its own flag: the manufactured misunderstanding — conflict one honest sentence would dissolve. Readers feel the author's thumb on the scale. If a conversation would fix everything, the story owes us a reason the conversation is impossible: pride with history behind it, a secret whose telling carries its own price. Otherwise the story owes us the conversation.
How to raise the stakes in a novel
Escalation gets misread as enlargement — more danger, a bigger explosion, a wider blast radius. But stakes aren't a quantity of peril; they're the price of losing as felt by this particular character, and raising them means narrowing: fewer options, less time, a higher and more personal price. Three dials govern it. Price — what losing costs. Proximity — who pays, with the dial moving from strangers toward loved ones toward the self. Pressure — how much room is left to maneuver.
In the house story, escalation never needs a villain. The buyer moves the deadline up. Daniel discovers that Mara has already signed a preliminary agreement. The buyer's plans, it turns out, include demolition. Each turn destroys a comfortable middle path that used to exist — sell later, split the proceeds, sell to someone who would keep the garden. That's the underlying logic: escalation is the systematic removal of compromise.
Two rules keep it honest. First, attempts must compound: when a character acts to repair the situation, the action should complicate it, or tension leaks out between chapters. Second, escalation by disclosure often outperforms escalation by event — the revelation that the stakes were always higher than anyone knew. That mechanism runs on setup and payoff, which is the plot twist's home ground, and it answers to the same rules of fair play.
Keep a ledger of wants
Most conflict problems on the page are bookkeeping problems underneath. The writer knew in chapter two what Daniel wanted; eighteen chapters later, he's agreeably helping with the sale because the plot needs him in the car, and no one in the room objects — least of all the author, who has forgotten the objection existed. So keep a ledger. For each principal character: what they want across the book, what they want in the current scene, what losing would cost them. Read it before drafting a scene. Amend it after, because outcomes change wants.
This is exactly the kind of memory work worth delegating. Project-aware AI novel writing software can hold the ledger as a living record rather than a document you forget to open. StoryProp works this way: decisions you confirm — who wants what, and at what cost — are promoted to records the agent reads before it drafts, and when a new passage contradicts one, the contradiction is surfaced for you to rule on instead of being quietly smoothed away. The wants stay opposed because they stay written down.
None of this is cruelty to characters, though conflict advice can sound that way. It's the opposite — taking each person's want seriously enough to let it collide with someone else's. Give everyone in the story something they can't have without cost. Then take away the option of waiting.
Friction holds only as long as you remember what everyone wants, and a long book outlasts anyone's memory. Tools like StoryProp keep those wants on the record and read them before drafting a line, so you direct the collisions instead of reconstructing them — useful for a story whose people have been too polite for too long.