How to write script coverage
Somebody handed you a stack of scripts and expects reports by Monday. Nobody explained what a good report looks like. This is the guide for that moment: how to read, what to write in each section, and how not to embarrass yourself with the verdict.
Understand what you're actually being asked to do
Script coverage exists so that someone more senior than you doesn't have to read the script. That's the whole job. Your report will be read instead of the screenplay, which means it has to do two things at once: represent the script accurately enough that nothing important is missed, and render a judgment clear enough that a decision can be made in two minutes.
You're also the safety net. If your company passes on something that becomes a hit, or champions something that falls apart in development, your coverage is the paper trail. That's not meant to scare you. It's meant to explain why specificity matters more than eloquence. Nobody has ever been burned by coverage that was too precise.
If you haven't seen the document you're about to write, start with a full script coverage example and template and keep it open while you work.
Read the whole script, and take the right notes
Read the entire screenplay in one sitting if you can. Coverage written from a fragmented read shows: pacing judgments especially require feeling the script in real time. While you read, keep running notes on four things:
Where the acts break
Note the page where the story commits to its premise and the page where the endgame starts. If you can't find them, that's not a failed note. That's a finding.
Every moment you were confused
Confusion is data. If you lost track of a character's motivation on page 40, an audience will too. Write down the page number in the moment, because you will not remember it later.
When you checked how many pages were left
The most honest pacing instrument that exists. Each time you catch yourself checking, note the page. Clusters of those notes are where the script drags, and they belong in your comments.
Lines and moments that worked
Coverage that's all weaknesses reads as lazy, and strengths are how execs find writers worth tracking even when the script is a pass. Capture the two or three moments that genuinely landed.
Write each section in the right order
The report reads top sheet first, but you should write synopsis first: summarizing the plot is the fastest way to find out whether the story actually tracks, and whatever breaks while you summarize becomes your best comment material.
Synopsis: one page, ending included
Present tense, chronological, no editorializing. Cover the protagonist's goal, the major turns, and the ending in full. The discipline is compression: a feature has 40+ scenes and you get about a page, so every sentence has to carry plot. Save your opinions for the comments.
Logline: written by you, not the writer
One or two sentences: protagonist, goal, conflict, stakes. Write it from what's on the page, not from the query letter. If the script's actual logline is worse than the writer's pitch of it, that gap is itself worth a comment.
Comments: specific or worthless
Cover premise, structure, character, and dialogue, one paragraph each, plus marketability if your company wants it. The test for every sentence: could this be pasted into coverage of a different script? “The dialogue is weak” fails that test. “From the midpoint on, characters state their feelings out loud; see the hospital scene on p. 62” passes. Page references are what separate professional coverage from a book report.
Top sheet: fill it in last
Title, writer, format, genre, budget feel, setting and period, draft date, your name, the date. Thirty seconds of work once everything else is done, and the first thing anyone sees, so get the writer's name spelled right.
Give the verdict like you mean it
Most coverage ends in Pass / Consider / Recommend, usually given separately for the script and for the writer. Here's what the words actually signal inside a company:
Pass is the default, and it's fine. Most scripts are passes. A pass with specific, respectful reasoning is a professional document, not an insult.
Consider is not the safe middle option, even though it looks like one. Consider means “I would defend this in a meeting.” If you consider everything to avoid being wrong, your coverage stops carrying information, and people notice fast.
Recommend stakes your name on it. Senior readers go years between recommends. If you're new and the script is genuinely exceptional, say so plainly and let the strength of your comments carry it.
Consider the writer, pass the script is the most useful verdict a reader can give, and the most common way new writers get tracked. Weak premise, strong voice: say exactly that.
Getting through the pile
Budget 2-4 hours per script when you're starting: up to two hours to read attentively and another one to two to write. It gets faster, mostly because the writing gets faster. The reading shouldn't: readers who skim get caught by the second act.
A workflow shift that's become common: generate a structured first draft with an AI coverage tool, then do your read and correct it. The machine handles the synopsis compression and the report scaffolding; you verify every claim against the script and add the judgment only a person in the building can add, like whether this fits what your company actually buys. Editing a draft is much faster than staring at a blank page, and your name is still the one on the verdict.
If you go that route with submitted material, the tool matters: OnDesk's AI script coverage produces the full report structure in minutes, never trains on uploaded scripts, and the first 3 reports are free. A personal chatbot account is the wrong place for someone else's unreleased script; we wrote an honest breakdown of where ChatGPT falls short for coverage if you're weighing it.