Guide
Native is grammar, not a filter you add at the end
Match the grammar of the feed rather than the look of a commercial. Vertical, one person speaking, captions on, the first frame already mid-thought, no title card, no logo sting. Native is a set of decisions made before you record, not a treatment applied afterwards. This page covers the seven signals a viewer reads first, the three endings that undo all of them, and the coverage ceiling. By the end you will know what to change on your next recording.
- Signals read before the first word
- 7
- Endings that undo all seven
- 3
- Cutaway ceiling at normal pace
- 45%
The viewer classifies person or company before they process a word
That judgement happens in under a second and it decides whether the script is ever heard. Nothing you write survives being classified as an advert at frame one.
The signals carrying it are all structural. Shape, framing, delivery, opening, text, sound and coverage. None of them is about resolution or colour.
That is why native cannot be applied at the end. A grade, a font or a filter changes none of the seven, because none of them lives in the finish.
Everything around your ad is somebody's holiday, somebody's dog, somebody's opinion. An ad arriving in that grammar gets a second of grace, and a second is what the hook needs.
The prize is not deception. It is the right to be heard by somebody who did not ask. That is the only thing an opening can win.
- Vertical, full frame, 1080 by 1920. A letterboxed file is an advert before anything else happens.
- One person, chest up, camera at eye height, an ordinary room visible behind them.
- Opening already mid-thought. No title card, no establishing shot, no logo.
- Spoken rather than performed, with a specific detail a script would have removed.
- Captions burned in, in the central band, because a muted viewer reads nothing you said.
- Cutaways that hide cuts and prove claims, held under the ceiling so the face stays present.
- Audio that sounds like a room rather than a booth. Recorded close, with nothing humming.
Three endings undo all seven, and one of them is in almost every ad
A logo sting. Thirty seconds of a person talking, then an animated brand mark. The viewer retroactively reclassifies everything they saw.
A produced call to action. A designed button graphic with a swipe animation is the grammar of a commercial. The ask is the moment you can least afford it.
Music that arrives from nowhere. A bed that was not in the room reads as post-production. Post-production reads as budget. Budget reads as advertisement.
The fix in all three is the same. Let the person ask, in their own voice, on camera. Then stop the video when they finish.
That feels unfinished the first time. It tests better than the alternative more often than anybody expects.
The same ad, two endings
Everything the first thirty seconds bought is spent in the last two. The right column costs nothing to produce.
| Undoes it | Keeps it | |
|---|---|---|
| How the ad closes | Animated logo sting | The person asks, then it stops |
| The call to action | Designed button graphic | Said out loud, once |
| Music | Bed arrives at the end | Room sound, or a bed from the start |
| What the viewer concludes | That was an advert | That was a person |
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Imperfection is tolerated. Deliberate roughness is a performance and it reads as one.
The format forgives a noisy room, a slightly wonky frame and a stumble over a word. It does not reward manufacturing those on purpose.
Scripted hesitations are the clearest example. Written filler reads as a performance of naturalness, which is more obviously false than a clean read.
The same goes for shaky camera work added afterwards, and for lowering quality to look authentic. Viewers grade authenticity harshly, and they grade it on content rather than on noise.
What reads as real is specificity. A number, a date, a name, a consequence, and the detail nobody would have invented.
The fridge humming in the background. The blown out window. The mug in shot with the wrong logo on it. Those are free, and they are the strongest evidence you have.
The test is simple and uncomfortable. Show the first two seconds to somebody who has not seen it and ask what they think it is. An answer of advert means the format decisions failed before the script was heard.
Coverage is where native ads quietly become montages
Cutaways make an ad feel edited. Past a point they make it feel produced. The person the viewer was following disappears under stock footage.
Above roughly half the runtime a talking-head ad is a stock montage with narration. That is the register people learned to skip years ago.
Cutroom enforces a ceiling by pace. 40 percent coverage at chill, 45 at normal, 52 at fast. Faster cutting carries more coverage before the effect breaks.
Three moments take none at all. The hook, because the viewer is deciding who is speaking. The proof, because the expression is the evidence. The ask, because it needs the viewer to be looked at.
Library footage is the other giveaway. The smiling team around a laptop and the anonymous city timelapse are read as filler by everybody, instantly.
How much of the runtime a cutaway may cover
The caps by pace. Above half the runtime, the person who made the ad feel native has left it.
- Chill pace40%
- Normal pace45%
- Fast pace52%
Seven signals, one recording session, and a file built in that grammar
Six of the seven signals assume a person on camera. That is the one input the format asks for, and twenty minutes covers several ads.
Cutroom is built in the same grammar. One talking-head take of up to three minutes returns a finished 9:16 MP4, directed on the transcript.
Highlight a phrase and a clip lands over exactly those words. Delete a line and the cut rebuilds around the gap. Change the pace and the whole read re-cuts.
The export carries no logo and no title card, because neither exists here. That is the ending the format wants.
Three facts to plan around. Three minutes of source. 9:16 output. Six caption packs, adjustable inside each, with no pack builder.
There is a route when nobody will film. A photo you supply plus a voice-over renders a lip-synced presenter at 320 credits per rendered minute. It carries an informational script well, and rarely carries first-person testimony.
Questions people ask
- Does native mean the ad should hide that it is an ad?
- No, and disclosure rules apply regardless of how the ad looks. Native means matching the format's grammar so the viewer gives it a second before deciding. The claim, the brand and any paid endorsement disclosure all stay explicit.
- Can a produced ad ever look native?
- Rarely, and the tell is usually the opening and the ending rather than the middle. A polished first frame and a logo sting classify the whole thing, whatever the twenty-eight seconds between them contain.
- Do I need to lower the video quality to look real?
- No. Deliberate degradation reads as a performance and is caught quickly. What reads as real is the room, the delivery and the specific detail. Keep the picture clean and spend the attention on the audio.
- Is a native look right for every placement?
- It is built for scrolling feeds. A placement the viewer cannot scroll past, or a product that sells on demonstrated craft, rewards production value instead. Those are the two cases where this page does not apply.
Seven signals, decided before you press record. Cutroom builds the whole ad in that grammar and hands back a 9:16 file with no logo sting to undo it.