A script is really just a sequence of scenes, and yet most craft conversations skip right past the scene itself, jumping straight from structure to dialogue to character, as if a scene were just a container those things get poured into. It isn't. A scene is its own unit of storytelling, with its own internal logic, and a script full of technically correct scenes that don't individually work is still a script that doesn't work.
Every scene needs a purpose: what changes by the end?
The single most useful question a writer can ask about any scene is also the simplest: what's different by the time it ends? Not what happened, what changed. A scene where two characters talk and nothing shifts, no new information, no altered relationship, no decision made, no stakes raised, is a scene that isn't earning its place on the page, no matter how well the dialogue reads in isolation.
Change doesn't have to be dramatic to count. It can be as small as a character admitting something to themselves for the first time, or a piece of information shifting who has power in a relationship. But it has to be real. A useful test: if you deleted this scene entirely, would the audience miss anything besides the words on the page? If the story would function identically without it, the scene is decoration, not structure.
Entering scenes late, leaving early
One of the clearest tells of an overwritten script is scenes that start before anything interesting happens and end after the interesting part is already over. A scene that opens with a character walking in, sitting down, and making small talk before finally getting to the point is asking an audience to wait for the actual scene to start.
The professional instinct is to enter as late as possible, right at or just before the moment of actual tension or change, and exit as early as possible, the instant that change has landed. Skip the greetings. Skip the goodbyes. Trust the audience to fill in the transitional connective tissue between scenes; they almost always can, and giving them that credit is what keeps a script moving at the pace the finished film is actually going to move.
This isn't about writing short scenes for their own sake. Some scenes genuinely need room to breathe. It's about making sure every beat inside a scene, from its first line to its last, is doing something, rather than easing an audience in and out of a moment they didn't need eased into.
Conflict and tension as the engine of scene work
Conflict doesn't mean characters yelling at each other. It means two things in a scene are pulling in different directions, whether that's two characters wanting incompatible outcomes, a character wanting something the situation won't allow, or even a character in conflict with themselves. Without some version of that friction, a scene has no engine, and a scene with no engine tends to just sit there, technically happening without actually building toward anything.
Tension is conflict's quieter cousin, and it's often more powerful precisely because it's less obvious. A scene where two characters are being carefully, deliberately polite to each other while something unspoken sits underneath every line can carry more dramatic weight than an outright argument, because the audience senses the gap between what's being said and what's actually happening. Learning to write tension without stating it directly is closely related to writing strong subtext in dialogue: the friction is felt, not announced.
A scene without conflict or tension isn't necessarily bad on a line-by-line level. It's just inert. It can be well-written and still fail to do the one job every scene actually has, which is to move something.
Visual storytelling vs. dialogue-heavy scenes
Screenwriting is a visual medium first, and it's worth remembering that a scene doesn't need dialogue to work, sometimes it works better without it. A scene told entirely through action, a character's face, a small gesture, the specific way someone enters or leaves a room, can carry as much emotional information as a page of conversation, and it does so in a way unique to film. The wordless standoff, the silent decision, the look exchanged across a room: these aren't gaps in the writing. They're often the writing at its most confident.
That said, dialogue-heavy scenes aren't a lesser form when they're earning their length. A tense negotiation, an argument that unravels a relationship, a confession, these scenes often need the specificity of spoken language to land, and trying to force them into pure visual storytelling can flatten what actually needs to be said aloud. The real skill isn't preferring one mode over the other, it's recognizing, scene by scene, which one the moment actually calls for, rather than defaulting to dialogue out of habit because it's the easier thing to write.
Diagnosing why a scene "isn't working"
Every writer eventually hits a scene that reads fine, technically, formatting correct, dialogue grammatical, and still feels dead. Diagnosing that feeling usually comes down to checking a short list of usual suspects.
Nothing changes. Revisit the purpose test above. If the scene doesn't move anything, forward, sideways, internally, that's very often the entire problem, no line-level polish will fix it.
The scene is solving a problem that isn't dramatically interesting. Two characters agreeing on logistics, working out a plan everyone already wants to execute, tends to be inert, because there's no real friction underneath it. If the scene's core question has an obvious, uncontested answer, there's no tension driving it.
It's entering too early or exiting too late. Trim the front and back half of the scene and see if it plays tighter and truer without the throat-clearing.
The conflict is external but not personal. A scene can technically have opposing goals and still feel flat if neither character actually cares about the outcome on a personal level. Raise what's actually at stake for the specific people in the room, not just the plot mechanics around them.
It's doing someone else's job. Sometimes a scene isn't working because it's actually two scenes awkwardly merged, or because it's carrying exposition that belongs somewhere else entirely. Isolating exactly what job a scene is supposed to be doing often reveals that it's currently doing several jobs badly instead of one job well.
Running a scene through this kind of checklist is exactly the sort of pass that's easier with a tool built for iteration rather than a static page. StoryWriter's structure makes it simple to isolate a single scene, test a trimmed version against the original, and see immediately whether cutting the first few lines or the last few actually tightens the read, without losing track of the draft you're comparing it against.
A scene is a story in miniature
Every scene, at its best, behaves like a tiny version of the larger script it lives inside: it opens with something unresolved, it complicates that thing, and it closes with something changed. Writers who internalize that scale, treating each individual scene with the same rigor they'd apply to the story as a whole, tend to produce scripts that never have a dead patch, because there isn't a single scene in them that was allowed to just sit there.