One unfakeable detail carries it. The rest is decoration.

Every viewer assumes a testimonial is selected. Most assume it might be paid.

That assumption is not fatal, because it can be overcome by detail.

I nearly returned it in the second week is a sentence nobody writes into a brief.

It carries the whole clip, because a brand would not have chosen to include it.

The same is true of physical detail. A slightly wrong angle. A dog walking through.

The bit where they lose the thread and come back to it.

Produced testimonials strip all of that out.

Better lighting, a tidier room, a script that removes the hesitation.

The result is indistinguishable from an actor, which is the one thing this format cannot survive.

So the production instruction is unusual. Collect the take with as little intervention as possible, and do the work in the edit.

Almost nobody has a testimonial shortage caused by unhappy customers.

They have one caused by never asking, or by asking in a way that is too much effort to answer.

The ask that works is narrow.

Not tell us what you think, which produces a paragraph of politeness.

Instead, what were you doing before you found this, and what changed in the first month.

Ask for a phone video under three minutes and say so explicitly.

The ceiling here is three minutes, and a shorter ask gets answered more often anyway.

Send the request within a week of the moment the product worked, not at the anniversary of the purchase.

Detail decays fast.

Collect several. The value is in having twenty to choose from, since the one usable sentence appears in maybe a third of them.

Make the request part of a process rather than a campaign, or it happens once and stops.

A raw customer video is mostly unusable, with one strong passage buried in the middle.

The job is removal, and removal is the thing a compiled cut does best.

Upload the take, up to three minutes.

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

Delete the lines where they warmed up and the cut rebuilds around what is left.

Trim the opening and the ending. Emphasise the phrase where they said the thing you would never have scripted.

Change the pace and it re-cuts the whole piece.

B-roll coverage caps at 40, 45 or 52 percent so the person stays on screen. In this format the face is the evidence.

Your product footage sits in the workspace library and lands on the words where they name it.

You can also fix a mis-heard word so the captions do not misquote your customer, which matters more here than anywhere else.

The usual pattern is one great testimonial run everywhere for a year.

It fatigues. It only ever speaks to viewers who see themselves in that one person.

Twenty testimonials is a different asset.

Different ages, different situations, different reasons for buying.

So the system has something to match against a wider audience.

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

Twenty is roughly 2,200 credits, which is one month on Premium at $79.99.

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

A set gives you a chance where a single favourite gives you a preference.

The trial is 7 days and 300 credits with no card, which is two finished cuts from footage already sitting in an inbox.

The ceiling is three minutes of source, so a long interview gets trimmed before upload or re-recorded as a short answer.

There is no colour work and no audio repair.

A customer filming in a loud kitchen produces a cut of a customer in a loud kitchen, so ask for a quiet room in the request.

Exports are 9:16 MP4 only and one project makes one finished cut.

Nothing here adds a disclosure badge or a verification mark of your own.

If a paid endorsement has to be labelled where you operate, say it out loud in the read and let the caption carry it.

Your product footage lives in the library and lands under the phrase where the customer names it.

This cannot manufacture a testimonial, and it should not.

A generated presenter reading a customer quote is not a testimonial, and viewers treat it as an advertisement.