How to Write a Short Video Script That Holds Attention

Why short scripts are harder than long ones

In a ten-minute video you have time to warm up. In a 30-second Short, every second competes with the swipe. There is no room for introductions, greetings or 'in this video I'll show you'. The script has to start in the middle of the interesting part.

The good news: short scripts follow a repeatable structure. Once you learn it, writing one takes minutes.

Count your words: about 20 per 8 seconds

A natural speaking pace is around 150 words per minute, which is about 2.5 words per second. That means an 8-second beat holds roughly 20 to 22 words. Go much higher and the voice sounds rushed; go lower and the visuals feel empty.

For a 32-second Short, plan four beats of about 20 words each, so 80 to 90 words in total. Write the script, read it aloud with a timer, and cut until it fits. Cutting almost always makes it better.

The hook: the first sentence decides everything

The hook is the first 2 to 3 seconds. Its only job is to make the viewer want the next sentence. Effective hooks usually fall into a few patterns:

• The surprising claim: 'This fort was attacked 12 times and never fell.'
• The direct question: 'Why does your phone battery die faster in winter?'
• The specific number: '5 minutes that change your whole morning.'
• The mistake: 'You're storing onions wrong.'
• The result first: 'This saved me an hour a day. Here's how.'

A four-beat structure for 32 seconds

Use this as a template and adjust the content, not the shape:

1. Beat 1 (0-8 s): the hook plus one line of context. Make a promise.
2. Beat 2 (8-16 s): the first half of the payoff. Give something real immediately.
3. Beat 3 (16-24 s): the second half, or the twist. Raise the stakes or add the surprising detail.
4. Beat 4 (24-32 s): the conclusion and a call to action, or a line that loops back to the start.

Write for the eye as well as the ear

Each beat needs a visual, so write a short scene note next to every line: what the viewer should see. 'Close-up of shoes being tied at sunrise' is far more useful than 'running scene'. Change the scene every beat at least.

Use short words and short sentences. Captions will show your words on screen, so a sentence that is hard to read is also hard to hear.

End with a reason to act

A call to action works best when it's specific and tied to the topic: 'Follow for part two, where we visit the fort that did fall' beats 'like and subscribe'. Another option is the loop ending: finish on a sentence that leads naturally into your first line, so the rewatch feels seamless.

A worked example

Here is a 32-second script on a fitness topic, written in four beats of about 20 words each:

1. Beat 1: 'Five minutes. That's all it takes to change how your whole morning feels. No gym, no equipment, just this.'
2. Beat 2: 'Minute one: stand up and drink a full glass of water. Minutes two and three: slow squats and arm circles.'
3. Beat 3: 'Minute four: step outside or open a window for daylight. It tells your body the day has started.'
4. Beat 4: 'Minute five: write one thing you'll finish today. Try it tomorrow, and follow for the evening version.'

Review before you render

Before producing anything, check the script against three questions: does the first sentence work on its own, is every fact correct, and does it fit the time? Rendering video costs time and money, while editing text costs nothing.

Cutflare's AI script writer builds this check in: one AI drafts the voiceover and scene notes in 8-second beats, and a second AI reviews the hook, the facts and the length before anything is rendered. It writes in English, Hindi or Hinglish styles, or you can paste your own script in any language.

Frequently asked questions

Related