Count the products you could talk about for forty seconds

A catalogue business cannot record a take per product. Nobody has three hundred takes in them and nobody should try.

Generating from a product page covers the long tail at a cost per asset that stays flat as the catalogue grows.

It also handles the cold start. A SKU that went live yesterday gets creative today, without waiting on a person to be free.

There is a staffing reality behind it. Filming three hundred takes is not a scheduling problem, it is a job nobody will take.

Past a certain SKU count the answer stops being a preference and becomes arithmetic.

Here is what a link cannot reach. A product page was written for somebody who already arrived.

An ad interrupts somebody who was not looking for you, and that needs a claim a person makes.

Cutroom scales on that instead. One spoken take goes in, up to three minutes.

Back comes a finished 9:16 MP4, captions burned in and coverage sitting on the words you marked.

One product, one audience, one offer means the variation has to come from angles rather than from more items.

Angles are claims: the objection you answer, the comparison you make, the moment of use you show, the reason you built it.

Every one of those is something a person says, and none of it is on the product page.

That is why a single-product brand runs out of link-generated variation fast.

One filmed take contains several claims. Marking different phrases and writing different opening headlines produces genuinely different ads from it.

Each of those exports costs 20 credits a finished minute, and the 100-credit batch is charged once for the take.

The symptom is easy to spot. Cost per result improves for a fortnight, then flattens, and new layouts stop moving it.

That is the point where the account needs a new claim rather than a new arrangement of the old one.

Upload the take. Cutroom transcribes it, drafts the whole cut and returns a 9:16 MP4 with captions burned in.

Highlight a phrase and a clip lands over exactly those words. Swap it, or search millions of free ones.

Delete a line and the cut rebuilds around the gap. Change the pace and the machine re-cuts the whole piece.

Emphasise the word your claim turns on so it pops in the captions. Fix a mis-heard word and the burned-in text follows.

Write a different opening headline, mark different phrases, export again. Same recording, a different ad.

No timeline, no keyframes and no layers. None of those decides whether a stranger keeps watching.

Coverage is capped by pace: 40 percent on chill, 45 on normal, 52 on fast. The person stays the spine of every version.

One talking-head take of up to three minutes goes in. One 9:16 MP4 comes out, captions burned in.

There is no link input, no catalogue import and no bulk generation.

A long product walkthrough cannot come in whole. Trim it or record something built to be an ad.

There is no square placement and no landscape master.

There is no product image library. Anything visual about the product lives in your take or in the stock search.

There is no timeline underneath, no bulk mode and no scheduler.

Those are the facts to plan around. What they do not touch is the cost of a second angle, which is next.

Most accounts are in the middle. Forty SKUs, three of which carry most of the money.

Cover the thirty-seven from links, where nobody films and the ads catch demand rather than create it.

Film the three. Those are the ones a cold audience has to be persuaded about, and persuasion needs a person.

Winners run at 5 to 8 percent of creatives, per Motion's analysis of 550,000+ Meta ads. Volume beats polish on both halves.

A second angle from a take already transcribed costs 20 credits and about four minutes.

One finished minute from a fresh take is about 120 credits, and every cut after it is 20 a minute.

Seven days and 300 credits with no card builds two ads for a hero SKU. Run them against the generated ones.