Guide
Thirty seconds is seventy-five words. Count them first.
A natural read is two and a half words a second. So a thirty-second ad is about seventy-five words. Most first drafts run a hundred and twenty. This page gives you the block budget, a worked script that fits it, and the two structural fixes that buy back five seconds. You will finish able to count a draft before anybody points a camera at it.
- Spoken words that fit inside thirty seconds
- 75
- Words in most people's first draft of the same ad
- 120
- Words in the hook, against words in the proof
- 8 / 22
The day of the shoot is a bad time to discover your draft is 120 words
A fifteen-second ad is around thirty-eight words. A sixty-second ad is around a hundred and fifty. Write your script, then count.
Almost everyone's first draft of a thirty-second ad is a hundred and twenty words. On the day there are two ways to fix that. Talk faster, or cut while the camera is running. Both are bad.
The budget forces the real decision, which is what to leave out. A short ad can carry one idea properly or three ideas badly.
Run a word count. Divide by two and a half. That is your runtime in seconds. It takes ten seconds, and it settles arguments that otherwise get settled by whoever is holding the phone.
Count your last three scripts before you write a fourth. Most teams find their thirty-second ads were forty-second ads read at speed. That is why the endings always sounded rushed.
The hook gets the fewest words and does the most work
This is a structure rather than a formula and the timings are starting points.
Notice where the words go. The proof gets the most, because a claim without evidence in a short ad is an assertion the viewer has no reason to accept.
Check your draft for the opposite shape. A long opening and a thin proof is the most common shape of a script that tests badly.
Count each block separately the first few times. Most people find their hook is fourteen words and their proof is six. That is the imbalance in one line.
- 0 to 3 seconds, the hook, about 8 words. One sentence whose only job is to buy the next three seconds.
- 3 to 8 seconds, the problem in their language, about 12 words. Say it the way a customer said it to you.
- 8 to 15 seconds, the mechanism, about 18 words. What the thing does and why that solves the problem.
- 15 to 24 seconds, the proof, about 22 words. A demonstration, a number, a before and after, or an objection answered.
- 24 to 30 seconds, the ask, about 15 words. One action, stated plainly, with the reason to do it now.
The word budget, block by block
Seventy-five words in total. The proof gets the most and the hook does the most work.
- Hook8 words
- Problem12 words
- Mechanism18 words
- Proof22 words
- Ask15 words
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
A worked script with no word a marketing department would have added
For a fictional product, a reusable coffee filter, to show the shape rather than to be copied.
Hook: 'I have thrown away about eleven hundred paper filters this year.' Eight words. A specific number, a first-person confession, and an obvious question raised.
Problem: 'Every morning, same routine, same bin, and the box runs out on a Sunday.' Twelve words in the customer's own frame, with no product yet.
Mechanism: 'This is steel mesh with holes fine enough to hold the grounds but not the oils.' Sixteen words describing an object rather than claiming a benefit. A viewer can picture steel mesh.
Proof: 'That is why the coffee tastes heavier. Paper absorbs the oils. I rinse this under the tap for five seconds.' Twenty-two words containing a real tradeoff and a demonstration.
Ask: 'One filter, about the cost of two boxes of paper. Link is below.' Thirteen words, one action, honest about price. Naming the cost removes the objection before the landing page has to fight it.
Thirty seconds, five blocks, in order
Every block has a job and a length. If any block is longer than its bar, another block is paying for it.
- Hook0 to 3 sec
- Problem in their words3 to 8 sec
- Mechanism8 to 15 sec
- Proof15 to 24 sec
- Ask24 to 30 sec
Two structural mistakes: a slow start and a stacked ending
Any script that opens by establishing context has spent the seconds meant to buy attention on scenery.
Move the most interesting sentence in your draft to position one. See what breaks. Usually nothing does. The context you thought was load-bearing was a throat-clear you wrote while working out what you meant.
The second mistake is a stacked ending. Some ads finish with the offer, the guarantee, the discount code and the urgency, all at once.
That hands the viewer a list to process at the exact moment they were supposed to act. One ask. The guarantee and the code belong on the page they land on.
Both fixes are deletions, and together they buy back about five seconds of a thirty-second ad.
Read the ending out loud with a timer running. A stacked close is easiest to spot by how long it takes to say.
The same script, before and after the two fixes
Both fixes are deletions. Neither costs anything, and together they buy back about five seconds.
| The draft that tests badly | The one that ships | |
|---|---|---|
| Opening | Three seconds establishing context | The most interesting sentence, first |
| Hook length | Fourteen words and a preamble | Eight words, said out loud first |
| Proof | Six words, mostly adjectives | Twenty-two words, one demonstration |
| Ending | Offer, guarantee, code and urgency | One ask, and the rest on the landing page |
One body, six openings, and every version exports for 20 credits
Get the structure right and the same body supports many openings. Record the openings back to back against one body. Now you have a hook test rather than a single ad.
That is the workflow Cutroom is built for. One take of up to three minutes becomes a finished 9:16 MP4. You direct it on the transcript, so a reordered script is a rebuild rather than a re-edit.
Delete the block that overran and the cut rebuilds around it. Trim the opening. Change the pace. Each version exports at 20 credits per output minute.
Nothing else prices the fifth version like the first. An editor quotes a reorder as a job. A timeline charges an evening for it.
The facts to plan around. The source is a talking-head take and somebody at your company records it. Three minutes is the upload ceiling. Every export is a 9:16 MP4, and there is no timeline underneath.
The counting stays yours. Seventy-five words is a decision made at a table with a pen, and it is the one that decides whether the ending sounds rushed.
Questions people ask
- Should short ads always follow this order?
- No. Leading with proof works well for visually demonstrable products, and leading with the objection works well in sceptical categories. The word budget per block is the durable part. The order is a starting point to deviate from deliberately.
- How long should a video ad be?
- As long as the argument needs and no longer. Shorter is not automatically better, and every extra second must earn itself. Write it at your natural length, count the words, then cut to the shortest version that still carries the proof.
- Do I need a call to action if the platform shows a button?
- A brief spoken one usually helps, because it tells the viewer what to expect on the other side. Keep it to one action and one sentence. Long spoken calls to action at the close of a short ad tend to reduce completion.
- How do I make the mechanism section not sound technical?
- Describe the physical thing plainly, the way you would to someone at a kitchen table. Steel mesh with fine holes, not advanced filtration technology. Plain nouns are more convincing than engineered language in a first-person ad.
- Who should ignore the word budget entirely?
- Anyone making an ad with no continuous speech in it. Silent product montages and motion-design pieces are counted in shots and beats, not words, and they need a motion tool rather than a transcript. The blocks still transfer. The arithmetic does not.
Seventy-five words for thirty seconds. Count them the night before, then let one take come back as a finished vertical ad the same morning.