Short-Form Video Script Structure: Hook, Setup, Value, CTA
The four-part structure behind short-form video scripts, with the word count each part should get by video length, three templates you can copy, and how each part fails.
If you are not on camera, the script is the video. There are no facial expressions to lean on and no charisma to paper over a weak line. The good news is that structure is learnable in an afternoon, unlike presence.
This is the structural reference: the four parts, how many words each one gets, three templates to copy, and the specific way each part fails. For the texture of the sentences themselves, see how to write scripts that sound like a person.
First, the word budget
Speaking pace is about 2.5 words per second, so length is arithmetic:
| Video length | Total words | Hook | Setup | Value | CTA |
|---|---|---|---|---|---|
| 15 seconds | 35 to 40 | 8 | 5 | 20 | 5 |
| 30 seconds | 70 to 80 | 10 | 12 | 48 | 8 |
| 45 seconds | 110 to 120 | 12 | 18 | 75 | 10 |
| 60 seconds | 145 to 155 | 12 | 22 | 105 | 12 |
Write to the count, not to the feeling. A 30-second video carrying 120 words is a 48-second video that people skip.
The proportions matter more than the exact numbers: the value section takes 60 to 70% of the words, and everything else is scaffolding around it.
Hook: the first three seconds
One line. It has to work with no setup at all, because there is none yet.
- “You are editing videos the slow way.”
- “I posted every day for 30 days without showing my face. Here is what happened.”
- “There is one setting that doubles your watch time.”
Four hook shapes that reliably work: a bold claim, a specific number, a question the viewer has already asked themselves, or the end of the story told before the beginning.
How it fails. It announces instead of starting. “Today I want to talk about video editing” is an announcement. Anything beginning with “Hey guys” has already lost. The other failure is a hook that needs context to make sense, which by definition it cannot have.
Setup: one or two lines
Just enough context for the value to land. Who this is for, or what the problem was.
- “I used to spend two hours per video trimming clips and syncing captions.”
How it fails. It becomes a preamble. Every sentence here is a sentence the viewer might leave during, so if the value would land without a line, delete the line.
Value: the middle, and the reason the video exists
Teach one thing. A method, a result, a story with a lesson.
- “Now I write the script first, paste it into a tool that generates the voice, the captions and the cuts, and I spend the time I saved on the next script.”
One idea per video. If you have three, you have three videos, and posting three is better for you than cramming three.
How it fails. Two ideas. It feels generous and it halves the completion rate, which is the metric that decides whether the video gets shown to anyone new.
CTA: the last line
Say exactly what to do. Follow, save, comment a word, open the link.
- “Save this for the next time you are staring at a blank timeline.”
How it fails. Asking for three things, or summarizing. The viewer was there thirty seconds ago and does not need a recap.
Together, those four parts come to about 75 words: a 30-second video.
Three templates you can copy
Replace the brackets, keep the rhythm.
The result template
“[Number] [thing] in [time] without [pain]. Here is what I did. First, [step one]. Then, [step two]. The part that mattered was [insight]. Do this and [result]. [CTA].”
The mistake template
“Stop [common action]. It looks productive and it is killing your [metric]. The reason: [mechanism in one line]. Do [alternative] instead. I switched [time ago] and [specific result]. [CTA].”
The story template
“[Moment where things went wrong]. I thought [assumption]. Then [turning point]. What I learned: [lesson in one sentence]. If you are [situation], [advice]. [CTA].”
When the structure changes
Under 15 seconds, drop the setup entirely. Hook, value, CTA. There is no room for context.
Over 90 seconds, the single value section becomes three beats with a small hook in front of each, because attention has to be re-earned roughly every 20 seconds.
Tutorials invert the proportions: the setup shrinks to nothing and the value section becomes numbered steps, where the numbering itself does the work of holding attention.
Using it
Write five hooks before you write anything else. Hooks are the hard part, and separating them keeps you from polishing paragraph two of a video that has no reason to exist. Then fill in the rest for the ones that survived.
When a script is finished, the production workflow takes it to a posted video. In MonkeyTalker that loop runs about three minutes per video once your character exists, and the free plan needs no card. Write one 75-word script with the structure above and put it through today.
Keep reading
Faceless Video Ideas: 10 Formats and What Each One Needs
Ten faceless video formats that work on TikTok, Reels and Shorts, with what each one demands of you in effort, voice and visuals, so you can pick the one you will actually finish.
Read the articleHow to Make Videos Without Showing Your Face
The full production workflow for faceless video: script length, choosing a voice, what goes on screen, captions and platform exports, from a written script to a posted video.
Read the articleHow to Write Video Scripts That Sound Like a Person
Why written scripts sound robotic when read aloud, and the editing habits that fix it: the read-aloud test, punctuation as pacing, and how to write for a synthetic voice without sounding like one.
Read the article