Guide
'I stopped setting the six-forty alarm' beats any production value
A viewer decides person or advertisement inside a second. Three signals settle it. A detail too boring to invent. Imperfections of delivery left in while genuine faults are fixed. A speaker who says in four words why they would know. This page works through all three, with the lines to keep and the lines to cut. Production value is not on the list, and by the end you will know where that budget should go instead.

- How long the person-or-advertisement decision takes
- 1 sec
- Odd concrete details to pull from your reviews today
- 5
- What stating your relationship to the product costs
- 4 words
Signal one: the detail nobody would bother to invent
Not 'it saves so much time' but 'I stopped setting the six-forty alarm'. Not 'great quality' but 'the handle is heavier than I expected and it does not rattle'.
Invented copy generalises. The writer does not have the details. Real experience is full of odd, useless-sounding particulars. The oddness is the point, because a made-up story would never include the boring bit.
Which is why mining your own support inbox and reviews outperforms writing. Customers hand you their specifics for free and in their own phrasing.
Go and get five of them now. The lid that sticks. The third week. The smell when you open the box.
Those five sentences beat a day of copywriting. Collecting them takes four minutes. Paste them somewhere you will see them before the next shoot.
The same claim, written twice
Every line on the right came out of a review or an inbox. Nobody could have invented the right column at all.
| What a copywriter writes | What a customer said | |
|---|---|---|
| Time | It saves so much time | I stopped setting the six-forty alarm |
| Quality | Premium build quality | The handle is heavier than I expected |
| Ease | Incredibly easy to use | I do two of the four steps and it still works |
| Result | Visible results fast | The third week is when my partner noticed |
Signal two: keep the imperfections of delivery, fix the imperfections of comprehension
A slightly uneven first sentence, a small self-correction, a laugh that starts before the line does. These read as a person talking. They are also the first casualties of a tidy edit.
A stumble that still communicates is an asset. A sentence nobody can follow is a defect. That distinction is the whole rule.
The same applies to picture. A real room with objects in it beats a styled background. Handheld with a slight settle beats a locked tripod.
Blown-out exposure, hollow echoing sound and an unreadable frame are not authenticity. They are faults. Viewers read them as low effort rather than high honesty.
And do not shoot rough on purpose. Deliberate degradation reads as a technique once an audience has seen it a hundred times. Unstyled and clear is the target.
- Keep: a stumble, a pause, an unpolished opening, a real background.
- Fix: bad audio, unreadable captions, a face lost in shadow.
- Avoid: styled sets, actor cadence, stock footage of aspirational strangers.
- Never fake: a testimonial nobody gave, or a result nobody got.
What a tidy edit removes, and what it should remove instead
The left column is the person. The right column is the frame around them. Most over-edited ads have these the wrong way round.
| Leave it in | Take it out | |
|---|---|---|
| The read | A stumble that still communicates | A sentence nobody can follow |
| The room | Objects, clutter, an actual kitchen | A hollow echo off hard walls |
| The camera | Handheld with a slight settle | A blown-out window behind the face |
| The timing | A laugh that starts before the line | Four seconds of dead air at the front |
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Signal three: who is talking, and why would they know this
If a viewer cannot answer both in the opening seconds, the ad is asking for trust it has not established.
Founders are plausible because they built the thing. Customers are plausible because they bought it. Staff are plausible because they use it daily.
A polished presenter with no stated relationship to the product is the least plausible option, and it is the one most brands default to.
State the relationship early and plainly. 'I make these.' 'I bought this in March.' Four words does more for credibility than any production value you could buy.
Say it before the claim rather than after it. Who is talking decides how the claim gets heard.
The corollary is a hard limit on synthetic presenters. A face with no relationship to the product cannot supply plausibility. Viewers are unusually good at detecting one performing a feeling it never had.
Polish the frame around the person, never the person
Real does not mean rough everywhere. The parts a viewer experiences as effort on their behalf should be clean.
Legible captions. Tight pacing with the dead air removed. No long silence at the front, and sound you can hear on a phone in a noisy room.
None of those signal a brand. They signal that somebody respected the viewer's time, which is the only production value a feed rewards.
Cutroom leans this way by design. It removes the dead time, times captions to your own audio, and puts footage on the exact words that need showing.
The voice, the room and the specifics stay yours, because you recorded the take. The stumble, the laugh, the odd detail about the third week, the kitchen behind you. That is the person, and nothing here touches it.
The frame gets tightened. The person comes through untouched.
The failure mode is easy to reach. If your finished ad could have been made about any product by any brand, the specificity has been edited out.
That is a decision about what to cut, and here it stays with you. The transcript is your own words. You delete the lines you decide are carrying nothing.
Everything the machine does sits around the person. Dead air removed. Captions timed to your own audio. Footage placed on the exact words that named something. One batch of 100 credits, one finished 9:16 MP4.
The facts to plan around. Somebody at your company is on camera, because a real speaker is the entire premise. Three minutes is the upload ceiling. Every export is a 9:16 MP4, and there is no timeline underneath.
Nothing of yours is stamped on the export either. On this page that reads closer to a feature than a limit.
It has no opinion about whether you sounded like a person. That judgement never leaves you, and it is the only part of this page that decides the ad.
Questions people ask
- Should the video look badly shot on purpose?
- No. Deliberate roughness reads as a technique once an audience has seen it a hundred times, and it is one of the fastest-ageing conventions in the format. Aim for unstyled and clear rather than degraded.
- Does the speaker need to be a real customer?
- They need to be plausible and honest. A staff member explaining how something works is credible without pretending to be a customer. What destroys trust is a scripted testimonial presented as a personal account.
- Do I have to disclose that an ad is an ad?
- Paid placements are labelled by the platform, and if anyone was compensated for their appearance you should say so on screen. Requirements differ by market, so check your local advertising standards.
- Can a generated presenter feel real?
- For informational scripts, often yes, because the speaker is functioning as a narrator. For first-person emotional claims, viewers are unusually good at detecting a face performing a feeling it does not have.
- Who should not chase this style at all?
- Brands whose credibility comes from institutional authority rather than a personal account, and anyone with nobody willing to appear on camera. Buy a studio production or book creators instead. A first-person ad with no plausible first person is the worst of both.
Specific, plausible, unstyled, clear. Your voice and your room go in, and the finished vertical ad comes back with every one of them intact.