How to write a short video script that holds attention
Most short videos fail on paper, not on camera. A five-part structure for 30-60 seconds, how many words fit, and how to write one in ten minutes.
A short video that loses people rarely loses them because of lighting. It loses them because the idea never arrives anywhere. Structure fixes that, and structure is decided before the camera turns on.
Write it backwards
The fastest way to write a script is in reverse order of delivery: closing line first, then the turn, then the hook.
A hook written before the payoff tends to promise something the video does not deliver, and viewers leave feeling misled even when the content was good.
- Write the one sentence the video exists for
- Decide what makes it different from the usual take on the topic
- Write the hook once you know where you are going
- Read it aloud against a stopwatch and cut what does not fit rather than speeding up
- Strip filler words: they reappear naturally when you speak
The five-part shape
- Hook: the first sentence answers why anyone should keep watching
- Problem: what is broken and why the viewer cares, in one sentence
- Turn: the part the viewer did not expect, and the reason they stay
- Payoff: how it works in practice, where the detail lives
- Line: one sentence a viewer can repeat to someone else
How many words fit
| Length | Words | What fits |
|---|---|---|
| 15-20 seconds | 40-55 | Hook and one claim |
| 30 seconds | 75-90 | Hook, problem, line |
| 45 seconds | 110-135 | The full five-part shape |
| 60-90 seconds | 150-230 | The shape plus a worked example |
Counting words at the draft stage saves a reshoot. Two hundred words in a thirty-second slot turns into a race; forty words in a minute leaves gaps nothing fills.
What breaks a script
- Three ideas instead of one, which produces a list nobody remembers
- Saving the conclusion for the end, school-essay style, when short video needs it audible in the hook
- Words nobody says out loud: any phrasing you would not use in conversation
- No pauses, which reads as an ad rather than a person talking
- A call to action longer than the point itself
Adjusting per platform
The same idea unfolds differently depending on where it lands. In a recommendation feed the viewer arrives cold and needs context inside the video. In a subscription feed they already know you, so the setup can be cut.
| Where | What changes |
|---|---|
| Shorts, Reels, TikTok | Hook is mandatory, context lives inside, repeat the line at the end |
| Newsletter or channel | Hook moves into the post text, the video can open on the point |
| Long form | Same shape, each part expanded with examples |
Where Monty fits
The script stays yours: idea, turn and closing line are not delegable. What automation removes is everything after you speak, which is pause cleanup, captions, per-platform cuts and posting. Fumbling a take stops mattering, because the fumble gets cut.
Bottom line
A good short script is one idea carried to a conclusion in forty-five seconds, with a turn in the middle and a hook that promises exactly what the video delivers. Written backwards it takes ten minutes, and it saves an hour of retakes.
FAQ
Does a forty-second video really need a script?
Short videos need it most. The less time you have, the more expensive the run-up is, and without a script the run-up eats the first three seconds where the decision to watch is made.
How many words per minute of video?
Roughly 150-230 words for 60-90 seconds, about 110-135 for 45 seconds. Counting at the draft stage prevents both racing and dead air.
Where should I start writing?
With the closing line. Then find the turn, and write the hook last, so it promises what the video actually delivers.
Will scripted delivery sound robotic?
Not if you strip filler and read it aloud before shooting. Naturalness lives in delivery, not in the absence of structure.
Sources
Hey Monty. Make me a reel.
Try Monty freeKeep reading
AI Reel Maker: Why It Should Start From Your Footage, Not a Template
AI reel makers split into template fillers and footage-first agents. Where templates work, why personal brands need real footage, and how one take becomes a posted reel.
Horizontal to Vertical Video: Smart Face-Tracking Crop, Not a Flat Center Crop
Turn 16:9 into 9:16 for Reels, Shorts and TikTok without chopping off heads. How smart face-tracking crop works, the real numbers, and safe zones.
Auto-Post to Telegram and Cross-Post Video Across Every Network
How to publish one video to Telegram, Reels, Shorts and TikTok at once. Cross-posting with no manual re-upload, where Telegram is a first-class platform.