How to Rewrite a Scene

There's a particular kind of scene that haunts every revision process. Not the scenes that are obviously broken and need to be cut. The scenes that are almost there. They do their job, the dialogue reads fine, the actors could perform them without complaint — but something keeps them from actually landing. You read them and think: this should work. Why doesn't it work?

I had one of those scenes in a script for almost eighteen months. It was a confrontation between two characters who'd been circling each other since the first act. The dialogue was sharp, the location was right, the emotional stakes were correctly identified. I rewrote it at least eight times before I finally saw what was wrong: both characters wanted the same thing. There was nothing to fight about at the level that mattered. I had created the appearance of conflict while carefully avoiding its substance.

That's the most common diagnosis for a scene that almost works. Here's how to find what's actually broken in yours — and how to fix it without rewriting forever.

The three diagnostic questions every scene has to answer

Before changing a single word in a troubled scene, run it through three questions. These aren't abstract principles — they're a checklist you can apply to a page of text in about two minutes.

Does the point-of-view character have a clear goal? Not a general mood or disposition — a specific, achievable thing they're actively trying to accomplish within the bounds of this scene. "Marcus wants to confront his father about the money" is a goal. "Marcus is angry and exhausted" is a state. Goals create forward momentum. States create atmosphere. If your scene has texture but no momentum, this is almost certainly where the problem lives.

Is there a genuine obstacle? An obstacle is whatever prevents the character from simply getting what they want and moving on. It can be another character with a competing agenda, an external circumstance that gets in the way, or a conflict between two things the protagonist needs simultaneously. If your scene resolves without resistance — the character wants something and gets it without friction — there was no scene. There was a transaction. Transactions don't hold audiences.

Does the scene turn? Something must change by the end. A relationship shifts. Information is revealed that reframes the situation. A decision is made that can't be taken back. A character doesn't get what they wanted, or gets it at a cost they didn't expect. A scene that ends with characters in the same position — emotionally, informationally, relationally — as when it began is a scene the story could survive without. If several scenes in a row are doing this, read why scenes feel flat — the underlying cause is usually consistent across all of them.

Why the conflict you're avoiding is usually the real scene

Here's something I've noticed over years of reading and writing scenes that don't quite work: they're often missing conflict because the writer is protecting the characters from it. Characters are kept polite, or at worst mildly frustrated, when the scene actually calls for genuine friction. The writer imagines a confrontation but writes a conversation.

This happens for understandable reasons. Writers develop real affection for their characters, especially the ones they've lived with across a whole draft. It feels somehow violent to put them through a real fight, to let them say something unforgivable or expose a wound that's been carefully protected. But scene after scene of tactful, reasonable characters is scene after scene with no dramatic energy. Conflict isn't cruelty — it's the engine that makes scenes run.

The practical fix: ask what each character in the scene actually wants, at the level beneath the stated subject of the conversation. Then ask whether those wants are in genuine opposition. If both characters want the same thing, rewrite one of them so they don't. A scene about choosing a restaurant can be a scene about control and contempt if the characters want different things at a deeper level than lunch. The surface conflict is just the vehicle for the real one — and the real one is what the audience feels.

Most scenes that almost work are missing conflict — not because the writer can't write it, but because they're unconsciously avoiding it.

When to try the opposite version of the scene

When a scene isn't working and I've run the three diagnostic questions without finding a clear answer, I use a technique I learned from a writer considerably better than me: write the exact opposite version of the scene.

If the scene is a quiet conversation, write it as a fight. If a character is being honest, make them lie. If it ends in reconciliation, let it end in a rupture. If the tone is tense and dark, try playing it as dark comedy. Write the opposite version completely — not as a sketch, but as a real draft of the scene. Ten minutes of work, maybe fifteen.

This exercise does two things. Sometimes it reveals a better version of the scene than what you had — an approach you wouldn't have considered without deliberately moving to the other extreme. And sometimes it shows you definitively why the original instinct was correct, which is equally useful because you stop second-guessing the approach and start focusing on what the execution was missing.

The key is committing to the opposite fully, not halfway. A scene that's "slightly more confrontational" doesn't produce diagnostic value. Going all the way to the opposite — making the scene as different from the original as possible — is what reveals what's actually available in the material.

Using subtext: what the scene says versus what it means

A scene can have a clear goal, genuine conflict, and a real turn — and still feel slightly flat if every character says exactly what they mean. Real human communication operates mostly beneath the surface of what's being said. When a screenplay lacks subtext, characters announce their intentions, explain their feelings, and state the theme directly. The scene is comprehensible but never quite alive.

Subtext in a scene means characters pursuing their real agenda beneath the surface of whatever the scene is ostensibly about. A scene about whether to sell the family house is really about whether one sibling thinks the other has earned the right to a say. A scene about what to cook for dinner is really about whether the marriage has a future. The characters never state the real thing directly. They handle the surface thing while the real thing accumulates in the space between their lines — and audiences feel that accumulation even when, especially when, none of it is named.

Adding subtext to a scene is a revision technique, not a first-draft quality. Here's how to do it deliberately: take the scene, identify what each character really wants and really feels (not what they'd say if asked), then rewrite the scene so those real wants and feelings are present but never spoken. The scene says one thing. It means another. The gap between the two is where the drama lives.

Line-by-line rewrite versus structural rewrite: knowing which one you need

There are two fundamentally different kinds of scene rewrites, and confusing them is how writers waste hours without improving anything. A line-by-line rewrite keeps the scene's structure and purpose intact while improving execution — sharpening dialogue, tightening description, clarifying individual beats. A structural rewrite means rethinking the scene from its foundation: who's in it, what they want, what happens, how it ends.

The mistake is doing a line-by-line rewrite on a scene that needs a structural one. You can spend two hours polishing dialogue in a scene that's broken at the level of goal and obstacle. The result will be well-written and still not work. You'll have lost those two hours and the scene will still be in the script.

How do you know which kind the scene needs? Run the three diagnostic questions first. If the scene fails on goal, obstacle, or turn, it needs a structural rewrite. If it passes all three but feels loose or slightly on-the-nose in execution, it needs a line-by-line pass. The diagnosis determines the treatment. Going straight to line-level work on a scene with structural problems is like refinishing a table with a broken leg.

Scene rewriting is detective work — you're looking for what it's missing, not what's wrong with what's already there.

When the scene is right but its position in the script is wrong

Occasionally, you'll address goal, obstacle, turn, and subtext across multiple rewrites, and the scene still doesn't quite land. The reason might have nothing to do with the scene itself. The scene might be structurally sound and simply displaced — in the wrong position relative to what the audience knows and feels when they arrive at it.

The emotional or informational context the audience needs for the scene to land might not have been established yet, or might have been established too far back to still be active. This is a structural problem, not a scene problem. The scene isn't broken — it's out of place. The fix might be moving the scene, adding a brief beat elsewhere that properly sets it up, or cutting the scene and distributing what it does across the surrounding ones.

If this kind of displacement is happening across multiple scenes, the issue is almost certainly in the larger architecture. See how to rewrite your entire screenplay for the process of approaching a structural overhaul from the foundation up.

The practical step-by-step when you sit down to fix a specific scene

Here's the actual process for approaching a scene you know isn't working:

• Step 1. Write one sentence describing what this scene must accomplish for the story — what must change, what the audience must know or feel by the end.
• Step 2. Run the three diagnostic questions: goal, obstacle, turn. Note which elements are absent or underpowered.
• Step 3. Write down each character's surface agenda (what they appear to want) and real agenda (what they actually want beneath the conversation).
• Step 4. If the scene fails the diagnostic, try writing the opposite version before starting the structural rewrite. Commit to it fully.
• Step 5. Rewrite the scene from scratch in a separate document, without looking at the original while you write. The original version can be a reference to return to — not a template to work around.
• Step 6. Read the new version cold the next day before deciding whether it works. Proximity distorts judgment in both directions.

The step most writers skip is Step 5 — writing from scratch instead of editing the existing version. Working inside the existing scene means you're operating in the mindset that produced the version that doesn't work. The scene's current shape subtly constrains every decision you make within it. Starting fresh, even when you end up keeping much of the same material, puts you back in an open creative state where better choices are genuinely available.

Scenes are the fundamental unit of screenwriting. Everything the audience experiences happens through them. A script composed of scenes that all work is, by definition, a script that works — which means individual scene rewriting is the most direct path to a better draft. The instinct to zoom out and address the whole script at once is understandable, but more often than not, the answer is in the room: find the scene that isn't landing, run the diagnostic, and fix what's specifically broken about it rather than what you imagine might be.

← All articles