They already know who you are. Skip the introduction.

The retargeting pool is people who visited, added to a basket, or watched most of a video.

They know the name, the category and roughly what you sell.

Then the ad opens by telling them the name, the category and roughly what you sell.

Three seconds spent on information they already have, at a higher cost per impression than cold traffic.

The advantage of this audience is that you can start in the middle of a conversation.

Nobody else gets to do that, and almost nobody uses it.

Start on the blocker instead. About the delivery time. About the size question.

About the reason it costs more than the one they were comparing it to.

That is a different script, not a different budget.

The reason it does not happen is production capacity. A second version for a smaller audience is hard to justify when a version costs a booking.

Someone who abandoned a basket at the shipping step has a different objection from someone who read the page twice and left.

The first is about cost or delivery.

The second is about whether the product suits them.

A single retargeting ad addresses at most one of those and annoys the other.

So the useful build is a small set, each one opening on a specific blocker.

Run them to the segments you can actually separate.

Price. Delivery. Fit or sizing. Whether it works with something they already own. What happens if it goes wrong.

Five ads, five sentences, one filming session.

Roughly 5 to 8 percent of creatives carry spend, per Motion's analysis of 550,000+ Meta ads.

A warm audience is small enough that fatigue arrives faster than that figure suggests.

A retargeting pool does not grow on its own.

It is fed by cold traffic and it is small, so the same person meets your ad repeatedly within days.

Two identical exposures work.

The fifth is worse than nothing.

The brand starts to feel like something following them around rather than something they were interested in.

The lever is rotation rather than budget.

Different openings, different faces, different answers, so the third exposure is a new sentence instead of the same one louder.

That only works if a replacement is cheap.

A finished thirty second ad is about 110 credits: 100 for the batch, 10 to render.

Basic is $39.99 for 2,500 credits, roughly 22 ads. Premium is $79.99 for 5,000, roughly 45.

Film one take of up to three minutes covering the product honestly.

Upload it and Cutroom transcribes the read, drafts the cut, places b-roll against the words and returns a finished 9:16 MP4 with captions burned in.

For each blocker, delete the general opening and let the cut rebuild around the sentence that answers the objection.

Highlight a phrase and a clip lands over exactly those words.

Change the pace and it re-cuts, with b-roll coverage capped at 40, 45 or 52 percent.

A face carries most of the runtime, and warm audiences respond to a person rather than to a montage.

Your product footage sits in the workspace library and lands on the words where you name the size or the part.

Fix a mis-heard word in the transcript so a price or a delivery window is captioned correctly.

That matters more on this audience than anywhere else, because they are checking a specific number.

Six caption packs plus colour, weight, size and position, with per-word emphasis on the answer.

This makes one finished vertical video per project.

It does not generate creative per product from a feed, which is how a large share of retargeting runs.

So treat it as the human layer alongside your catalogue ads.

That layer is the part a feed cannot produce, and it is the part a hesitating visitor responds to.

Exports are 9:16 MP4 only. No square and no widescreen.

Source takes cap at three minutes and one project makes one finished ad.

The free b-roll search covers the world around the answer, and your own uploads sit beside it in the library.

Nothing of yours is stamped on the render, so the caption treatment carries the brand.

Somebody has to be on camera. If nobody will, the avatar module builds a presenter from a photo and a voice-over at 320 credits per minute of render.