Guide
Measure it as a share of the video. Not as a number of clips
The right ceiling for cutaway footage is a percentage of runtime, somewhere between 40 and 52 percent depending on how fast the edit moves, and not the clip counts most guides give you. A rule like three to seven seconds per clip tells you nothing about whether your viewer has seen a human face in the last fifteen seconds, which is the thing that decides whether a talking head ad works.
- Coverage ceiling at a calm pace
- 40%
- Coverage ceiling at a normal pace
- 45%
- Coverage ceiling at a fast pace
- 52%
A clip count tells you nothing about what the viewer has actually seen
Most advice on this is phrased as clip length or refresh rate. Change the picture every four to eight seconds, keep each cutaway between three and seven.
Follow that on a thirty second ad and you can end up with the speaker on screen for eight seconds in total, which is a slideshow with a voice over it rather than a person selling something.
The same rule on a ninety second video is far too little coverage and the edit drags. The advice is not wrong so much as measuring the wrong quantity.
Share of runtime is the quantity that transfers. It is the same instruction at fifteen seconds and at three minutes, and it directly answers the question the viewer is unconsciously asking, which is whether there is a person here.
A thirty second ad at the normal ceiling
Roughly thirteen seconds of cutaway across the whole ad. The rest is the person making the argument.
Where the ceiling sits, and why it moves with pace
At a calm pace we cap coverage at 40 percent of runtime. At a normal pace it is 45, and at a fast pace it is 52.
The reason it rises with speed is that fast cutting shortens every individual cutaway. Six clips of a second and a half is a very different experience from two clips of four and a half seconds, even though both are nine seconds of coverage.
Short cutaways read as emphasis. Long ones read as a change of subject, and a change of subject is what pulls a viewer out of a direct address.
Above roughly half the runtime something flips. The stock footage stops illustrating the argument and starts being the video, and the person becomes a voice over somebody else's pictures.
That matters commercially because the face is doing work no clip can do. Trust in a short ad is carried by a person appearing to mean what they say, and you cannot buy that on a stock library.
Coverage ceiling by pace, as a share of runtime
Faster edits are allowed more because their cutaways are shorter. The ceiling is about attention, not about the clips.
- Calm40%
- Normal45%
- Fast52%
- Where it stops being your ad60% and up
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
What too little looks like, because that failure is real too
A ceiling is not a target. An ad with no cutaways at all is perfectly viable and plenty of the best performing direct response video is one person and a phone.
Coverage earns its place in two situations. When you name a thing the viewer cannot picture, and when the read has a flat stretch that needs a visual change to survive.
Everywhere else a cutaway is costing you the face for no return. If you cannot say what a clip is doing, cut it and give the seconds back to the person.
The honest test is to watch with the sound off. If you cannot tell who is talking or what they are selling, you have too much coverage. If it is visually monotonous for fifteen seconds at a stretch, you have too little.
When this rule inverts completely
Product demonstration is the obvious case. If the sale depends on watching the thing work, the footage is the ad and the talking head is the interruption.
Anything showing a process, a transformation or a place has the same shape. A tour is not a talking head video with heavy coverage, it is a different format.
Tutorials invert it too. On a how-to, the screen or the hands are the content and the presenter is a voice with occasional appearances.
Cutroom is built for the first shape and not these. It assumes somebody is speaking and treats footage as coverage over that read, so if the pictures are the point, a timeline editor will serve you better.
Questions people ask
- Is there a right number of b-roll clips for a 30 second ad?
- There is a right total duration, which at a normal pace is around thirteen seconds. How many clips that becomes depends on your pace, and the count is a consequence rather than a target.
- Does more b-roll improve retention?
- Up to a point, then it reverses. Visual change holds attention and losing the speaker loses trust, and the crossover in a direct address ad sits around half the runtime.
- What if I have no footage to cut away to?
- Then do not cut away. One person talking straight to camera for thirty seconds is a completely legitimate ad, and it beats a cutaway that does not match what is being said.
- Does Cutroom let me go above the ceiling?
- Not automatically. The compiler caps coverage at the ceiling for the pace you chose. You can add or swap clips yourself in the editor, so the cap is a default rather than a wall.
Count the seconds, not the clips, and give the rest back to the person making the argument. The pace you pick sets that ceiling automatically here, which is one fewer thing to eyeball on every cut.