How to Edit Reaction Videos

A reaction video has one asset: a face changing. Everything else in the frame is support. Most reaction edits fail because they spend their time on the thing being reacted to and put the face in a corner.

The format works because it is social. The viewer is watching someone else watch something, which is a way of watching it with company. That means the edit has to preserve the sequence of the reaction, not just the reaction itself.

The structure that works

Three parts, and the proportions matter more than the content.

1. The setup (2 to 4 seconds). Enough of the source clip that the viewer knows what is about to happen. Not the whole thing.
2. The reaction (60 to 70 percent of the runtime). The face, in real time, including the parts where nothing happens.
3. The payoff (5 to 10 seconds). The moment the reaction peaks, held, then a line that lands it.

The mistake is inverting the proportions: forty seconds of source footage with a small face in the corner is a repost with commentary, and it performs like one.

Framing two things at once

Vertical framing makes this harder than it looks, because you have one narrow column and two subjects.

• Split screen, face on top. The most legible layout for phone viewing, because the face is at eye level and the source is below. Keep the split at roughly 40/60.
• Face full frame, source inset. Better when the reaction is the point and the source is context. The inset should be large enough to read expressions and small enough not to compete.
• Cut between them. The most cinematic option and the most work: hold on the face, cut to the source at the moment of the reveal, cut back to the face for the aftermath.

Whichever you choose, keep the layout identical for the whole clip. A layout that changes mid-video resets the viewer's attention.

The cut points are the joke

Reaction editing is mostly about where you cut, not what effects you add.

• Cut before the peak, not after. If the reaction is a laugh, cut to the face half a second before it starts, so the viewer sees it arrive.
• Hold through the pause. The pause before a reaction is the tension. Cutting it out to keep pace kills the format.
• Cut back for the double take. A second look at the face after the source continues is the cheapest reliable laugh in the format.
• Do not cut on every beat of the source. The source clip's rhythm is not your clip's rhythm; the reaction sets it.

Sound is the whole trick

Reaction audio is two sources in one track, and getting the balance wrong ruins both.

• Mic the reactor, not the room. The reaction needs to be close and dry; the source clip can be thin and distant, because that is how it would sound in the room.
• Duck the source when the reactor speaks. A side-chain or a manual dip, either way: the voice has to win.
• Keep the source audible. Muting it entirely removes the reason the face is reacting, and the viewer loses the context that makes the reaction meaningful.
• Watch for the headphone bleed. Reactors wearing headphones often have the source leaking into the mic, which makes the whole clip sound like it was recorded in a bathroom.

Authenticity and the appearance of it

The format's currency is that the reaction is real. That has practical editing consequences: over-cutting, added sound effects, and zoom punches on every beat all read as manufactured, and manufactured reactions do not travel.

The edits that help authenticity: fewer cuts, longer holds, a visible pause, and no music under the reaction itself unless the reactor is genuinely reacting to music.

There is also an honest limits note worth stating: reacting to someone else's video uses their footage. Short reactions with visible commentary are widely tolerated and often licensed; long unedited stretches with a face in the corner are not, and they are the ones that get struck. Keep source clips short, keep your commentary load-bearing, and do not build a channel that only works if nobody enforces copyright.

Batching reactions

Reaction content scales unusually well because the recording is cheap: one camera, one mic, one sitting.

1. Pick five source clips the day before and watch them cold on camera, one take each. The first take is the reaction; the second is a performance.
2. Cut each one the same day, while you remember which moment surprised you.
3. Keep one caption preset and one layout for the whole batch: caption readability.
4. Publish across the week rather than in one day: publishing and scheduling.

The format also benefits from a consistent opening frame, because viewers recognise the layout before they read the title. That consistency is worth more than any individual edit.

When the format stops working

Three signals that a reaction channel needs a change of shape:

• The reaction is the same every time. If your face does the same thing for every source, the viewer has seen it and will not return.
• You are reacting to things nobody knows. The format depends on shared context; niche sources shrink the audience.
• The comments are about the source, not about you. That means you are a distribution channel for someone else's work, and you have no moat.

The fix in all three cases is the same: bring something the source does not have. Expertise, a contrary take, a demonstration, a story the source reminds you of. The face is the format; the reason to watch has to be something else.