In brief: A reel loses people when the first line promises nothing. YouTube’s retention report names four moments worth reading: flat stretches, gradual declines, spikes and dips. Write the promise first, keep one idea per text card, and if the reel is a paid post, write the disclosure label into the script before the edit begins.
Reel script writing is a structure problem before it is a creative one. Most reels lose people because nothing was promised in the first line, so there was no reason to stay for the second.
A script fixes that. It does not have to be long, and for a 20 second reel it fits on half a page. What it must do is decide what you promise, what you show, and what you want the viewer to do at the end. If you are planning a month of these, our guide to a social media content calendar covers the wider schedule.
Key Takeaways
- Open with a promise, so the viewer knows what they get.
- YouTube names 4 retention moments: flat stretches, gradual declines, spikes and dips.
- Spikes mean people rewatched that part, so study them.
- A Shorts view now counts on play, with no minimum watch time.
- On a paid reel, the disclosure label belongs in the script.
What Does Reel Script Writing Actually Involve?
Five parts, and you can write them on a phone in ten minutes.
A promise in the first line, then a reason to believe it. After that come two or three beats that actually deliver what you promised, a payoff the viewer can see, and one clear instruction about what to do next.
Most brand reels skip straight to the beats, which is why they feel like a slideshow with music. The viewer never learned what they were about to get, so they had nothing to wait for.
Write the promise last if that helps. Shoot the useful part first, then work out what single line would make somebody stay for it, and put that line at the top. Plenty of good reels are built backwards this way.
How Do You Write the First Line?
Say the thing, and say it to one person.
Vague openings die because they ask the viewer to guess. “Tips for your home” promises nothing. “Your kitchen chimney is why the whole flat smells” promises something specific, and it names who it is for.
Three openings that carry weight: a problem the viewer recognises, a result they want, or a claim they doubt. All three work because each creates a small gap that only watching can close.
Say it in eight words or fewer, because longer lines get read while the thumb is already moving. The discipline is tightest on a paid piece, where you are buying the attention rather than earning it, and our ad reels page shows the formats we build to.
What Does Holding Attention Look Like in the Data?
There is a real report for this, and it names four moments.
YouTube’s key moments for audience retention report “explains how well different moments of your video held viewers’ attention” (YouTube Help). It marks four things, and each tells a scriptwriter something different.
A flat line means “viewers are watching that part of your video from start to finish”. Gradual declines mean “viewers are losing interest over time”. Spikes appear “when more viewers are watching, rewatching, or sharing those parts”. Dips mean “viewers are abandoning or skipping at that specific part”.
Read your dips before writing the next script, because they point at the sentence that lost people. A dip usually sits where you explained something nobody had asked about yet.
Spikes matter as much. If people rewatched a two second stretch, that stretch is the reel, and your next script should start closer to it.
One caveat: the report is YouTube’s, and its intro figure measures who was still watching after the first 30 seconds, which is longer than most reels run. The shape of the chart still shows you where interest fell away.
Why View Counts Will Not Tell You This
Because a view no longer means somebody watched.
YouTube changed how Shorts views are counted from 31 March 2025. Views now “count the number of times a Short starts to play or replay, with no minimum watch time requirement” (Get started creating YouTube Shorts).
So a big view number can sit on top of a reel nobody finished. Judge a script on retention, rewatches and saves instead, and treat views as a reach figure.
That also means your first line is doing more work than the number suggests. It is the only part guaranteed to be seen.
Instagram reports its own set of numbers inside the app, and they are not the same ones. Pick one measure of holding power, whether that is watch time or rewatches, and judge every script by the same one for a few months.
How Long Should the Reel Be?
As long as the promise takes to pay off, and then you stop.
YouTube’s own tools now make Shorts “up to 3 minutes long”, so length is now a choice you make. Most brand reels do their job in 15 to 30 seconds, because one promise rarely needs more.
The useful test is per line. If a line neither sets up the promise nor pays it off, cut it. Cutting is what turns a 40 second script into a 22 second reel that people finish. The line you are most attached to is usually the one to lose.
For what a produced one costs once the script is right, our note on reel production cost in Ahmedabad sets out the ranges.
How Do You Test Two Hooks?
By changing one line and nothing else, which is the only way to learn anything from the result.
Shoot the body of the reel once, then cut two versions with different opening lines and post them a week apart to the same audience. Keep the thumbnail, length, music and caption identical, because every extra variable turns a clean test into a guess.
Give each version the same number of days before you judge it, and compare where people dropped rather than how many turned up. A hook that holds people ten seconds longer is worth more than one that pulled a bigger crowd and lost them immediately.
Write down what you learn in one line, somewhere the next scriptwriter will see it. Teams lose this knowledge constantly, because the person who ran the test moves on and the finding was never written down.
How Do You Write for Sound Off?
Assume the words have to work on screen, then treat audio as a bonus.
Put the promise on screen as text in the first frame, in a size that survives a small phone. Keep captions on every spoken line. If the reel makes sense with the sound muted, it will also make sense with it on.
Do not lean on automatic captions to do this for you. YouTube says its automatic captions are generated by machine learning and “might misrepresent the spoken content due to mispronunciations, accents, dialects, or background noise”, and it asks creators to review and edit them (Use automatic captioning). For Indian brand names and Gujarati words, that warning is not theoretical.
Two habits help here. Write the on-screen text in the script itself, so it is planned before the shoot. And keep each text card to one idea, because two ideas on screen means neither gets read.
Where Should the Brand Appear?
Later than most brand managers want, and earlier than the viewer notices.
Opening on your logo spends the one moment you cannot get back. Open on the problem or the result, and let the brand arrive when the viewer already wants to know who is talking.
The exception is a product the reel is actually about. If the thing on screen is the thing you sell, show it working in the first two seconds, because that is the promise doing its job. A logo sitting still does none of that work.
One note for anyone with brand guidelines to satisfy. A small mark in the corner usually keeps the brand team happy without spending the opening moment.
Our guide to reels for travel brands shows how that plays out in one vertical.
What About the Call to Action?
One instruction, at the end, matched to what the reel just did.
A reel that taught something earns a save. A reel that showed a product earns a link tap or a visit. Asking for a follow after a product demo is a mismatch, and mismatched asks get ignored.
Say it in plain words, and make the benefit obvious in the same breath. “Save this for your next site visit” beats “check out our page for more amazing content”, because the first tells somebody exactly what to do and exactly when it will be useful to them.
Put the ask on screen as well as in the voice. Losing somebody at the last second because they could not hear it is a poor way to end.
If It Is a Paid Reel, the Label Belongs in the Script
Because the timing is fixed, and adding it later ruins a cut.
For a branded post, ASCI requires a disclosure label on screen. On a video of 15 seconds or less it must stay for at least 3 seconds. On videos longer than 15 seconds and under 2 minutes, it should stay for one third of the length (ASCI influencer guidelines).
Write that into the script as a line, alongside the on-screen text. An editor who learns about the label after the cut is locked will either squeeze it somewhere it cannot be read, or reopen the edit. Neither is a good use of a shoot day.
Our influencer marketing team writes those briefs with the label built in from the first draft.
A Sample 20-Second Script
Here is an example, and the business in it is invented.
Notice what it does. The promise lands in the first line, the beats deliver it, and the ask matches what the viewer just learned. Nothing in it needs a studio.
No label is needed here, because it is the brand’s own reel. On a paid creator post the label would run from the start for about seven seconds, a third of twenty.
What Do Most Reel Scripts Get Wrong?
Scripts we are asked to rescue share the same five problems.
Fixing the first two usually changes the retention chart on its own.
If you would rather hand this over, our reels and video production team scripts, shoots and edits in house. Talk to us and we will read your last three scripts before you shoot another.
Frequently Asked Questions
How long should a reel script be on paper?
For a 20 second reel, roughly 45 to 60 spoken words at a normal speaking pace, plus your on-screen text. If it runs past a page, the reel is trying to do two jobs and should be two reels.
Do I need a script for a simple product clip?
Yes, even if it is four lines. The script is where you decide the promise and the ask, and those two decisions separate a clip that sells from one that just exists.
Should the hook be spoken or written on screen?
Both, wherever you can. The text carries it for anyone watching muted, and the voice carries it for everyone else. Our post on growing an Instagram account covers posting these consistently.
How do I know if my hook worked?
Look at where people leave. YouTube’s retention report marks dips as the points where viewers are abandoning or skipping, so a dip in the first seconds is a hook problem.
How many scripts should we write before shooting?
Write three and shoot the best one. The first is usually the obvious angle, the second a variation on it, and the third is where most teams find the line that actually says something.
Who should write the script, us or the agency?
Whoever knows the customer’s objection best. The strongest scripts we produce start from a sentence a client’s sales team says every day.
