How to Add B-Roll to Clips Without Burying the Speaker

The dead middle

Most talking-head clips lose people in the same place: seconds 12 through 25. The hook worked, the payoff has not arrived, and there is nothing to look at but a face that has been in the same position for twenty seconds.

B-roll is the cheapest fix. A cutaway to relevant footage resets visual attention without touching the audio, so the viewer keeps hearing the point while their eye gets something new. Done well, it buys you the middle of the clip.

Done badly, it does something worse than nothing: it makes the clip feel like a stock-footage montage where the speaker is incidental, and viewers leave because it reads as filler. The difference is entirely about whether the cutaway relates to what is being said at that exact moment.

Placement, duration, and the two rules

Where it goes. Over concrete nouns and specific references. Someone mentions a city, a product, a stadium, a chart — that is a cutaway. Over abstract statements ("I think the industry is changing"), B-roll has nothing to illustrate and becomes decoration.

Never over the first three seconds. The opening is where the viewer decides. A face and a clear caption line is the strongest opener you have. Cutting to footage before you have established who is talking makes the clip feel like an ad.

Never over the payoff. The punchline, the number, the reveal — the viewer wants to watch the face when it lands. Covering that moment is the most common B-roll mistake and it costs you the reaction shot that made the clip worth cutting.

How long. Two to four seconds. Under two reads as a glitch. Over five and the viewer forgets there is a person talking, which is the point of no return for a talking-head clip.

How much. Roughly 20-30% of the clip's runtime, so two or three cutaways in a 30-second clip. Above 50% you have made a different kind of video, and it needs to be intentional if you go there.

The two rules that cover most of it: relate to the words being spoken right now, and cut back to the face before the emotional beat.

Getting it done at volume

Sourcing B-roll manually is the bottleneck — finding a clip, checking it is licensed, trimming it, placing it. Ten minutes a cutaway adds up fast when you post daily.

AI B-roll on Pro and above handles the sourcing and placement: relevant footage matched to what is being said and inserted at points that do not step on the hook or the payoff. You review and adjust in the timeline editor rather than assembling from scratch. Pro is $39.99/mo and also carries background music, AI sound effects, spoken hooks, and caption translation and dubbing in 31 languages.

If you are on Starter at $19.99/mo, B-roll is not included, and that is often fine — plenty of high-performing clip channels use none at all. Reaction-heavy and comedy content usually plays better without cutaways, because covering the face removes the thing people came for.

What to check on any clip with B-roll before it goes out: does each cutaway relate to the line under it, does the face come back before the payoff, and does the whole thing still make sense with sound off? That last one catches most problems. See pricing for what each plan includes, and advanced B-roll techniques for the manual craft version.

When to use none

Cases where adding B-roll makes the clip worse:

Emotional or vulnerable moments. The face is the content. Cutting away breaks the connection you were clipping in the first place.

[Reaction content](/use-cases/reactions). The reaction is the point. This should be obvious and it still gets violated constantly.

Fast comedy. Punchlines land on delivery. A cutaway in the wrong half-second kills a joke that would have worked.

Anything under 15 seconds. There is no dead middle to fill. Just let it run.

Content that already has visuals. Gameplay, screen shares, demonstrations — you already have something to look at.

The clean test: watch the clip with no B-roll first. If your attention drifts somewhere in the middle, that is where a cutaway goes. If it does not drift, you do not need one, and adding some anyway is how competent clips become generic ones. Completion rate is where you will see whether a change helped, and it is the only judge worth listening to. Punch-in zoom is the alternative fix for a dead middle when cutting away is wrong — how to use punch-in zoom covers that.