How to Write a Screenplay: The First-Time Writer's Roadmap

The coverage came back on a Tuesday. Two pages, single-spaced, from a reader I had paid ninety dollars and would never meet. My logline scored "consider." My structure scored "consider." My dialogue scored "pass," and underneath that word, in a box labeled Additional Comments, was a paragraph that began: "The writer clearly loves these characters more than the audience will get the chance to." I read it standing up in my kitchen, then sat down, then read it two more times looking for the part where she was wrong.

She was not wrong.

The script was ninety pages of two people I adored talking about their feelings in a diner booth, and nothing in the world was making either of them get up from it.

That was my first spec script, and the coverage taught me more in two pages than four screenwriting books had taught me in four months. It taught me that a screenplay is not a story you tell. It is a story you build out of things people do, on camera, in an order that creates pressure. Everything else -- the voice, the theme, the beautiful line you have been saving since page one -- only survives if the structure underneath it is doing its job.

This is the guide I wish that coverage had come with. Not a formula, because good screenplays resist formula the way good jokes resist explanation, but a working process: premise, logline, outline, the beats a first draft genuinely cannot survive without, the brutal importance of your first ten pages, why film is a visual medium in ways prose is not, the format rules nobody explains properly, how to revise once the draft exists, what coverage actually means when it lands in your inbox, and where a finished script goes next. I am going to name the beats professionals use, because pretending structure does not exist helps nobody, and I am going to tell you where structure stops mattering, because that happens too, usually right around the point where a reader starts calling your script "alive."

Start With a Premise That Can Survive an Elevator

Before you open a blank page in Final Draft or anywhere else, you need a premise you can say out loud in one breath and have a stranger lean forward. Not a genre ("a heist thriller"). Not a setting ("set in 1920s Chicago"). A situation with pressure built into it: a person, a want, and a force actively working against that want.

Jordan Peele's premise for Get Out is a single sentence that does all the work a premise needs to do: a Black man meets his white girlfriend's parents for the first time, and something underneath their hospitality is very wrong. You do not need to know the third act to feel the movie in that sentence. The premise for Bong Joon-ho and Han Jin-won's Parasite is almost as compact -- a poor family cons its way, member by member, into working for a wealthy one -- and the whole film is an escalation of the pressure that single situation creates. Neither premise needed a twist to be gripping. The twist is what happens once the pressure has been running long enough to demand one.

Here is a test I use now, learned the hard way: if your premise needs backstory to make sense, it is not a premise yet, it is a setup for one. "A woman discovers her late mother kept a second family" is a premise. "A woman who grew up an only child, whose mother died when she was twenty-six, whose father remarried when she was thirty" is a biography, and biographies do not have engines. Cut until you find the sentence with an engine in it.

A useful second test, once the sentence has an engine: can you say what the protagonist loses if the story goes wrong for them. Not "conflict." Loss. Peele's protagonist can lose his freedom and, the movie eventually makes literal, his own body. Bong and Han's family can lose the fragile foothold they have just built, which is worse than never having had it. If you cannot name the specific loss, your premise has tension in theory and none on the page, and readers can tell the difference in the first three pages whether or not they could articulate why.

I run every new idea through both tests now before I let myself get attached to it, because attachment arrives fast and is hard to walk back once a dozen pages exist. Most premises fail the second test quietly -- there is a want, there is an obstacle, but nothing is actually at risk, and the resulting script reads like competent people solving a manageable problem. Manageable problems do not sustain ninety minutes. Specific, named losses do, and naming the loss before you outline a single scene will save you months you would otherwise spend discovering the same thing on page eighty.

The Logline Is Not Marketing, It Is Load-Bearing

Once you have a premise, compress it into a logline: one or two sentences naming your protagonist (a trait, not a name), the inciting event, the goal, and the stakes. Writers treat the logline as a chore for the query letter later. That is backward. The logline is the single clearest description of your movie's engine, and if you cannot write one, the engine is not built yet, no matter how many pages you have drafted.

The poet Marie Howe has said that a poem should work like a photograph and not an essay -- a single frame the reader can see, not an explanation of what the frame means. A logline needs exactly that discipline, and it is a strange comfort that a note about compression from outside the industry entirely lands harder than most screenwriting advice about loglines does. Your logline is not summarizing your movie. It is showing the reader one frame -- protagonist, want, obstacle -- clear enough that they can see the other ninety pages implied inside it.

I keep my logline taped above the monitor while I outline, and every beat I add has to answer to it. Does this scene serve the goal in the logline. Does it raise or lower the stakes named in the logline. If a scene cannot answer either question, it is probably backstory wearing a scene's clothes, and it goes in a separate document, not the outline. A structured logline template walks through the same four elements -- protagonist trait, inciting event, goal, stakes -- if you want a place to draft and stress-test a handful of versions before committing to one.

One warning from experience: do not fall in love with a logline that is more interesting than the movie you can actually write. I spent three months on a premise I could describe brilliantly at parties and never once outlined past page thirty, because the logline promised a story my actual interests could not sustain for ninety pages. A logline you cannot stop talking about and a logline you can finish are not always the same sentence. Write several. Pressure-test each one against your own attention span before you pick.

Building the Outline Before You Trust Yourself With Dialogue

Outlining a feature is its own discipline, and I am not going to re-walk every beat here -- the full outlining guide covers the beat sheet, the forty-card method, and scene-by-scene breakdowns in the depth they deserve. What I want to say here is smaller and, I think, more useful: outline further than feels necessary, because dialogue is seductive and it will happily carry you thirty pages in a direction your structure cannot support.

Vince Gilligan has talked about breaking every season of Breaking Bad in a room with a whiteboard before a single scene got written, because television punishes structural drift across thirteen episodes the way film punishes it across ninety minutes -- badly, and often too late to fix cheaply. Aaron Sorkin, whose dialogue is often cited as the reason his scripts work, has said just as often that he does not sit down to write a scene until he knows exactly what it needs to accomplish structurally. The dialogue people praise him for is downstream of an outline decision, not a replacement for one.

I resisted this for years because outlining felt like the opposite of the reason I wanted to write scripts in the first place, which was to hear characters talk. It took a second failed draft to understand that the outline is what protects the talking. Without one, every scene has to simultaneously advance plot, reveal character, and entertain, and dialogue written under that much pressure usually collapses into exposition dressed up in a voice. With a solid outline underneath it, a scene already knows its job before you sit down, which frees the dialogue to do something more interesting than carry information -- it can carry subtext, contradiction, the thing a character will not quite say.

Five Beats a First Draft Cannot Survive Without

Whatever beat-sheet model you outline with underneath it -- and there are several good ones -- a workable minimum for a first feature comes down to five markers you should know before you write a line of dialogue: the opening image, the inciting incident, the midpoint reversal, the lowest point, and the ending. Everything between those five can and should surprise you in the draft. Those five themselves should not be discoveries you make on page seventy.

The opening image sets tone before a word of dialogue explains anything. The inciting incident should land by roughly page twelve in a hundred-and-ten-page script, and it needs to be external and concrete -- something that happens to the protagonist, not a feeling they have. The midpoint reversal is where the story's rules change: a false victory that will curdle, or a false defeat that is actually the beginning of real strength, and either way the protagonist's understanding of their own situation has to shift, not just the plot's circumstances. The lowest point costs the protagonist something real -- Bong and Han's family in Parasite loses its foothold catastrophically and irreversibly at almost exactly this point in the film, and the cellar sequence that follows earns every minute of dread because the loss that preceded it was concrete, not implied. The ending mirrors the opening image closely enough that the distance between the two is legible as the story's argument.

This is not a cage. It is closer to a keel -- the thing that keeps a boat with a lot of sail on it from tipping over the first time the wind shifts. Charlie Kaufman's Adaptation. spends its middle section actively interrogating whether these five markers are even honest, and it is one of the most structurally daring American screenplays of the last three decades precisely because Kaufman knew the five markers well enough to bend them on purpose. You cannot subvert a structure you have not learned. You can only accidentally omit one, and a reader will feel the omission even when they cannot name it.

Test the five markers against your own outline the way you would test a bridge: not by admiring it, but by asking where it would fail under load. Read only your opening image and your ending, back to back, with nothing in between. Do they rhyme, or contradict each other, or simply fail to speak to each other at all? Then read only your inciting incident and your lowest point. Does the lowest point cost the protagonist something that the inciting incident's promise made possible, or did it arrive from a direction the first act never pointed toward? A structure that survives both of these spot checks will usually survive a reader's full attention. One that fails them will fail in the same place a stranger's eye lands first, which is worse, because you will not be in the room to explain what you meant.

The First Ten Pages Decide Whether Anyone Reads Page Eleven

Readers, agents, and contest judges read the first ten pages of your screenplay with genuine attention and the remaining hundred with whatever attention you earned in those first ten. This is not fair, and it is also not going to change, so build your first ten pages knowing it.

What those pages need to establish: who this story is about, what world they live in, what tone the movie is going to hold for ninety minutes, and the promise of the central conflict, even if the formal inciting incident lands a page or two later. Kaufman's Adaptation. breaks nearly every convention about pacing and voice in its opening pages and gets away with it because the tone -- anxious, self-lacerating, formally inventive -- is unmistakable within the first two. You do not need to imitate his approach. You need to be that decisive about what your own two pages are promising.

Damien Chazelle's Whiplash opens on a single drummer, alone, practicing, before a voice we cannot yet place interrupts him. No dialogue explains the stakes. The scene itself is the promise: this movie is going to be about pressure applied to a person in a room, and it delivers on exactly that promise for the next hundred and five minutes. That is what a first scene is for. Not exposition. A binding promise about what kind of experience the reader is about to have, one the rest of the script then has to honor.

Practically: cut anything in your first ten pages that exists only to orient the reader. A character explaining their own backstory to someone who already knows it is the single most common first-act mistake I read in other writers' drafts, and I have committed it at least four times in my own. If two characters both already know a fact, they do not say it to each other. Find another way in, even if that way in is harder to write than the exposition would have been.

Writing to Camera Instead of Narrating the Page

A novel can tell you what a character is thinking. A screenplay cannot, and the writers who struggle longest with the form are almost always the ones who came to it from prose and keep reaching for a tool the medium does not give them. Everything in a screenplay has to be something a camera could photograph or a microphone could record. "She remembers her mother" is not a scene direction. It is a wish, dressed as one.

This constraint is not a limitation to work around. It is the entire craft. Chazelle's opening drummer tells you everything about obsession and isolation without a single line about what he is feeling, because the visual choice -- alone, in a small room, repeating the same phrase until it is close to right -- does the work interiority would have done in a novel. Peele's dinner-table scenes in Get Out build dread almost entirely through blocking and reaction shots implied on the page: who is looking at whom, who looks away first, what a smile is doing that the words underneath it are not.

Practically, this means action lines should describe behavior specific enough to imply psychology without stating it. Not "John is nervous" but "John straightens the cutlery that is already straight." The second version gives an actor and a director something to do. The first gives them an adjective. When you find yourself writing a character's internal state directly, stop and ask what that character would actually do with their hands, their eyes, their posture, if a camera happened to be running. The answer is almost always more interesting than the adjective was, and it is the only version of the moment a viewer will actually receive.

Format Rules That Exist for Reasons, Even When They Feel Arbitrary

Screenplay format looks fussy to first-time writers and it is not. It is an industry-wide agreement that one formatted page equals roughly one minute of screen time, and everyone from your director to your line producer relies on that ratio to budget, schedule, and cast the thing you wrote. Break format badly enough and the page-to-minute math stops working, and experienced readers notice within the first page, before they have read a word of your actual story.

The core rules, in the order a first-time writer usually needs them. Twelve-point Courier, because it is a monospaced font that makes the page-per-minute estimate reliable in a way proportional fonts do not. Scene headings in capitals -- INT. or EXT., the location, the time of day, as in INT. DINER -- NIGHT. Action lines in present tense, third person, describing only what a camera could see or a microphone could hear -- no interiority, only what plays on screen. Character names in capitals above their dialogue, centered on the page. Parentheticals used sparingly, for tone or action that genuinely cannot wait for the next action line, never as a shortcut for dialogue that is not doing its job on its own.

William Goldman, in Adventures in the Screen Trade, wrote that "screenplays are structure," and one thing that observation misses on a first read is that format is structure's smallest visible unit. A page that is eighty percent unbroken action block with no white space reads, correctly, as a slow scene before anyone gets to the content. A page of clipped, one-line paragraphs reads as fast before a single word of dialogue lands. Learn to read your own pages the way a reader will -- for shape, at a glance, before content -- and format stops feeling like bureaucracy and starts feeling like another instrument you get to play.

A smaller point that costs first drafts more than it should: consistency matters more than any individual rule. If you call a character MARGARET on page two and MAGGIE on page thirty with no scene explaining the shift, a reader loses a half-second re-orienting every time, and half-seconds compound across a hundred pages into genuine fatigue. The same goes for scene heading conventions -- pick DAY or CONTINUOUS or LATER and use it the same way every time it recurs. None of this is creative. All of it is professional, and readers notice the absence of it faster than they notice its presence.

A few smaller conventions trip up first drafts specifically. V.O. after a character's name means voice-over, a line heard but not spoken on screen by a body present in the scene -- narration, a letter being read, a phone call where you are staying on the listener. O.S. means off-screen: the character is physically present in the scene's location but not currently in frame. CONT'D after a character's name, or at the top of a new page, signals that the same character is still speaking across an interruption or a page break, and Final Draft and similar software will usually add it for you. None of these earn their place through cleverness. They earn it by telling a director, in one abbreviation, exactly where to point a camera and where not to, which is the entire point of the format existing at all.

Writing the Draft Without Editing the Draft

Once the outline is solid, write fast. This is the one piece of advice from craft books that turned out to be completely true for me: a screenplay draft written in six focused weeks is almost always better than the same screenplay written across eight scattered months, because momentum carries tonal and structural decisions forward in a way that stop-start writing cannot.

You will want to fix page fourteen while you are writing page sixty. Do not. Make a note -- one line, in a separate document -- and keep moving. The version of you who revises is a different, more useful version of you than the version who is still discovering what happens next, and mixing those two jobs slows both of them down. Phoebe Waller-Bridge has described writing early material for Fleabag for stand-up first, at speed, discovering the character's voice through live performance before the structural work of adapting it for television began -- two distinct jobs, done in order, not simultaneously.

Finish the draft before you show it to anyone, including yourself in a more critical mood. A screenplay with an unfinished third act is not a screenplay with a good first act. It is an unfinished screenplay, and no amount of polish on page one changes that fact for a reader who reaches page ninety and finds nothing there. I have three screenplays in a drawer that stopped at page seventy because I kept going back to perfect the opening instead of finishing the ending, and every one of them is worse for it -- not because the endings would have been bad, but because I never found out.

The Revision Pass Nobody Wants to Do First

Revise structure before you revise dialogue. I know this backward. Everyone knows this backward the first time, because dialogue problems are the ones you can hear while structural problems require you to hold the whole shape of the script in your head at once, which is harder and less immediately satisfying.

Do a structural pass first: read straight through, ideally aloud or with cards laid out by scene, and ask only whether every sequence is pulling weight. Does the midpoint actually reverse something. Does the protagonist's decision at the act break come from agency or from the plot needing them somewhere at that exact moment. Fix what breaks at this level before you touch a single line of dialogue, because dialogue polish on a scene that gets cut in the next pass was never revision. It was avoidance dressed as productivity.

Your antagonist deserves a structural pass of their own, and it is the one first-time writers skip most often. Give them a plan that exists independently of the protagonist -- their own goal, their own timeline, their own idea of what winning looks like -- and check every scene they appear in against that plan rather than against what the protagonist needs from them narratively. An antagonist who only reacts to the hero is not a character. They are a wall with a face painted on it, and readers can tell the difference between a wall and a person by around page thirty.

Only after structure holds do you drop into line-level work: trimming action description to what a camera actually needs, cutting dialogue that says what a look could say instead, checking that every character's voice is distinct enough that you could cover the names and still know who is talking. That last test is unforgiving and worth doing anyway. I ran it on the diner script, after the fact, and discovered that both leads sounded like me on a good day. That was the real note underneath the coverage, before the reader ever put it into words.

Reading Coverage Without Either Panicking or Dismissing It

Coverage is the written evaluation a script reader produces for a production company, agency, or contest -- usually a logline, a synopsis, and a verdict of some version of Pass, Consider, or Recommend, plus comments explaining the grade. Almost every script gets a Pass on its way through the system, including scripts that eventually sell, so a single Pass is information, not a verdict on your future.

Read coverage twice. The first time, for the feeling -- let yourself be stung, because pretending it does not sting is its own kind of dishonesty and delays the second, more useful reading. The second time, read only for patterns. One reader saying your second act sags is an opinion. Three readers independently saying the same thing, in different words, across different submissions, is a diagnosis, and diagnoses are gifts even when they arrive rudely.

The coverage that changed my diner script did not tell me to add a car chase or raise the stakes with a bigger secret. It told me, correctly, that I had confused affection for my characters with dramatic tension between them, and that the two are not the same thing and never will be. I have carried that distinction into every script since. It was worth the ninety dollars roughly forty times over, and I say that as someone who resented every one of those dollars on the Tuesday I paid for it.

Where a Finished Script Actually Goes

A finished, revised, coverage-tested script has three realistic paths, and most working screenwriters pursue more than one at once. Fellowships and labs -- the Nicholl Fellowship, the Austin Film Festival competition, the Sundance labs -- read blind, which means a script with no connections and a strong logline genuinely competes on the page. These are slow, expect months rather than weeks, but they are the closest thing the industry has to a meritocratic front door, and placing well in one is a credential that opens actual meetings.

Querying agents and managers directly is faster in theory and far less reliable in practice. A cold query lives or dies on your logline and your first ten pages doing exactly the work described above, because that is often all an assistant reads before deciding whether the material moves upstairs. Festivals built around screenwriting specifically -- as opposed to finished films -- exist to create the rooms where readers, managers, and other writers actually talk to each other, which matters more than most first-time writers expect, because this remains an industry that runs substantially on people vouching for people they have read.

Table reads, formal or informal, belong somewhere on every one of these paths and most writers skip them out of nerves. Hearing actors, even amateur ones, stumble over a line you thought was clean will tell you more in twenty minutes than another silent read-through will tell you in a week. Dialogue that looks fine on the page and turns out to be unspeakable out loud is one of the most common gaps between a script that reads well and a script that actually works, and the only reliable way to find that gap is to put the words in someone else's mouth before a stranger does it for you, professionally, with money involved.

None of these paths are fast, and I do not think faking urgency about the timeline helps anyone. What helps is having more than one finished, revised script when the first one starts circulating, because the question every one of these paths eventually asks is not "did you write one good script" but "are you someone who is going to keep writing them." Answer that question with your second script while your first one is still out in the world.

My diner script never sold. It sits in a folder I have not opened in three years. But it taught me to read a room of two characters for pressure instead of affection, and that lesson shows up, uncredited, in every script I have finished since -- which is, in the end, closer to what that first coverage letter was actually trying to hand me than the word "pass" at the top of the page ever let on.

Plan your screenplay end to end in one workspace -- documents for your outline and treatment, a beat sheet you can rearrange as the structure changes, and a visual plot grid for tracking acts and sequences at a glance -- with Plotiar. Free to start.