Script writing for short videos: structure that keeps people watching
A short video script structure that holds attention: hook, setup, beats and payoff, with timing, a two-column template, sample scripts and mistakes to avoid.
The short answer
A short video script that keeps people watching has four parts: a hook that creates a question, a quick setup, a few beats that each move the answer forward, and a payoff that delivers what the hook promised. Every line either builds toward the payoff or gets cut.
Write it in two columns, what the viewer hears and what the viewer sees, so the words and pictures work together instead of repeating each other.
The structure
- Hook (first two to three seconds): a line and image that make the viewer want the answer. A surprising claim, a visible problem, a bold promise you will keep.
- Setup (a few seconds): the context the viewer needs, nothing more.
- Beats (most of the video): two to four steps, each with its own visual. Each beat should feel like progress.
- Payoff: the answer, result or twist the hook promised.
- Close: one clear next step, or a loop back to the start for rewatches.
The hook makes a promise. The payoff keeps it. If your payoff does not match your hook, viewers feel tricked and leave earlier next time.
Use the two-column format
Write audio on the left and visuals on the right. This stops two common problems: narration that describes exactly what is on screen, and visuals that do not match the words.
Good pairing: audio says "Most people store tomatoes in the fridge," while the visual shows a hand pulling a pale tomato from a cold shelf. Weak pairing: audio says "Here is a tomato," while the visual shows a tomato. The visual should add information, emotion or proof the words cannot.
Sample script: a 30-second tip
A kitchen tips account. Topic: why tomatoes taste better on the counter.
- Hook. Audio: "Your fridge is ruining your tomatoes." Visual: a tomato sliced open, pale and mealy, slow push-in.
- Setup. Audio: "Cold changes the texture and dulls the flavor." Visual: a fridge door closing on a bowl of tomatoes, cool blue light.
- Beat 1. Audio: "Keep ripe ones on the counter, stem side down." Visual: tomatoes on a wooden board, warm window light.
- Beat 2. Audio: "Eat them within a few days." Visual: a hand picking one up and gently pressing it.
- Beat 3. Audio: "If you have to chill them, let them warm up before eating." Visual: a tomato on a plate next to a kitchen timer.
- Payoff. Audio: "Same tomato. Completely different taste." Visual: split screen, pale slice beside a deep red, juicy slice.
- Close. Audio: "Follow for one kitchen fix a day." Visual: the board with the tomatoes, caption on screen.
Read it aloud with a timer. If it runs long, cut words, not beats.
Write for the ear
Short video scripts are heard, not read. Use short sentences. Put the important word at the end of the sentence. Avoid long numbers and lists of more than three. Use "you" and talk to one person.
Cut throat-clearing. Lines like "Hey guys, in this video I'm going to show you" waste the most valuable seconds you have. Start with the hook.
Write captions from the final script, then trim them. Captions should match what is said, but they can drop filler words to stay readable.
Pacing and timing
Each beat needs a new visual. If the picture stays the same for too long, attention drifts even when the words are good. In AI video, that usually means a new shot or a camera move for every beat.
Match length to the story, not to a target. A tip that needs twenty seconds should not be stretched to sixty. Platforms have different length sweet spots, and some programs have length rules: as of October 2026, TikTok's Creator Rewards only counts videos over one minute. Check the platform's current page when length matters for monetization.
In ClickPop, the YouTube Shorts studio at /app/youtube-shorts is built around this structure: hook, beats, captions and payoff in a vertical frame.
Three structures to try
Hook, beats and payoff is the base. These three variations fit most topics.
- Problem and fix: show the problem, show the mistake people make, show the fix, show the result. Good for tips and product videos.
- List with a twist: three items, where the last one is the surprise. "Three things to check before you rent an apartment. The third one costs people the most." Keep the strongest item for last.
- Mini story: a character wants something, something gets in the way, they find a way through. Works for drama shorts and brand stories with an original character.
Whichever you pick, write the payoff first. Once you know where the video lands, writing the hook and beats that lead there is much easier.
Mistakes to avoid
- A hook that promises something the video never delivers.
- Explaining the setup for so long the viewer leaves before the first beat.
- Narration that just describes the picture.
- Several ideas in one short. Make several shorts instead.
- Ending without a clear next step.
- Writing claims about health, money or results you cannot back up.
Questions
How long should a short video script be?
As long as the idea needs. Read it aloud and time it. Cut any line that does not move toward the payoff.
What makes a good hook?
A line and image that open a question the viewer wants answered, which the video then answers. Avoid hooks you do not pay off.
Should I write the script before generating AI video?
Yes. A script tells you which shots you need, so you generate only what the video uses and spend fewer credits.