CapCut makes the first ad. Cutroom makes the ninth

The auto-captions are accurate. The template library tracks what is working on the platform this month, which most editing software gives up on after a year.

It runs on the phone in your pocket. It works on a train. Somebody who has never edited anything can post something watchable the same afternoon.

What CapCut charges is published on their own site and it moves, so read it there on the day you decide.

The job changes shape somewhere around ad eight. One product video a fortnight is a craft task. Twenty ads a month across six SKUs is a production line.

A timeline is built for the craft task. Every ad starts from an empty track. Every decision passes through your hands, one at a time.

That is fine at ad one. At ad eight on a Thursday night it is the reason ad nine never gets made. Cutroom is built so that ad nine costs what ad one did.

Run one product for a quarter and you will build about twenty ads for it. The offer stays put. The hook changes. The cutaways change.

On a timeline everything else changes too, quietly. Caption size drifts two points. The opening title sits half a second longer. The music comes in earlier. The cut before the price is tighter on ad four than on ad eleven.

Nobody decided any of that. One person made the same forty decisions twenty times and got tired somewhere in the middle.

This is where it costs money rather than pride. A test only tells you something when one variable moved. Ad four and ad eleven differ in the hook, plus eleven small ways you did not intend. The winner does not tell you what won.

You spent the budget. You waited the four days. The read is unusable. Then you do it again next month.

Published benchmarks put the share of creatives that become winners at roughly 5 to 8 percent, from Motion's analysis of 550,000+ Meta ads. At that rate a muddy test is not a small waste. It is most of your month.

Record one take of up to three minutes on a phone. Upload it. Spend 100 credits on a batch.

What comes back is already cut. Transcribed, trimmed, covered with footage, captioned and paced. Not a template with your clips dropped in. A cut built from what you actually said.

You change it by touching words. Highlight a phrase and a clip lands over exactly those words. Delete a line you do not want and the cut rebuilds around the gap. Change the pace and the whole read re-cuts.

There is no timeline underneath. No keyframes, no layers, no effects browser. That is the trade and it is what buys the speed. The machine owns every time calculation, so it can rebuild a cut in seconds.

Consistency arrives as a side effect, because the same compiler builds every export. Cutaways never exceed 40 percent of the runtime at chill pace, 45 at normal and 52 at fast. Captions come from six shipped packs. Transitions are Cut, Whip, Punch, Glitch and Sweep.

So ad four and ad eleven differ in the hook, because the hook is the only thing you changed. That is the whole point of running the test.

A batch is 100 credits. That covers the transcript, the director pass and the b-roll search for one take.

Export is 20 credits per output minute, so a thirty second ad costs 10 to export.

Twenty ads from ten takes is 1,000 credits of batches and 200 of exports. 1,200 credits, against the 2,500 Basic carries at $39.99 a month.

Premium is $79.99 for 5,000 credits. A top-up is $15 for 1,000 when a month runs long, which is the month a product launches.

The trial is 7 days and 300 credits with no card. Two takes and four finished thirty second ads is 240 credits, so the whole comparison fits inside the free week.

Credits are not the constraint on this route. The one input it needs is somebody willing to stand in front of a phone for twenty minutes.

The change that matters is not the software. It is when you record.

Stop filming per ad. Set aside one morning. Put the phone at eye height near a window and read one continuous take per product, under three minutes each. Six SKUs is under half an hour.

Ramble slightly at the edges while you are there. The extra sentences become options later, since deleting a line is one gesture rather than a reshoot.

Then the ads come out of that morning for the rest of the month. New hook, same take, new cut. The product footage you already own lands on the words that name it.

The behaviour change operators notice first is that they stop rationing ideas. The angle nobody was sure about gets made anyway, because making it costs a batch and an export instead of an evening.

That is what twenty ads a month buys. More of them, alike where you meant them to be alike, and consistent enough that a test finally reads.