Layout is the first thing a viewer judges and the last thing most reaction creators think about. Before anyone hears a word you say, they have already decided where to look. Get that wrong and your best take lands in a corner nobody was watching. Get it right and the video does half its work before you open your mouth.
There are only three layouts that matter, and the correct answer depends less on taste than on one question: where is this clip going to be watched?
The three layouts, honestly compared
1. Picture in picture: your face in a corner
The classic. Footage fills the frame, your camera sits in a small box in a corner, usually bottom right. It is the layout most people picture when they hear "reaction video," and it has a real strength: the source material stays big. If the whole point of your video is a spectacle shot, a monster reveal, or a fight sequence that deserves the full frame, picture in picture protects it.
The cost is that your face is small, and small faces do not carry emotion on a phone screen. A raised eyebrow at 15 percent frame width is invisible. The layout also puts you literally in the corner of your own video, which is a strange place to be if the reason people subscribe is you.
Use it when the footage is doing the heavy lifting and your job is punctuation. A trailer built on scale, like The Odyssey with its widescreen compositions, loses something real when you crop it down to make room for yourself.
2. Split screen: side by side, equal weight
Two panels, footage and face, sharing the frame. On a desktop or TV this is the most readable option there is. Nothing is hidden, nothing competes, and the viewer's eye can move between the two without anything shrinking.
The problem is geometry. Two 16:9 panels side by side on a horizontal canvas means both are now tiny. Split screen was designed for widescreen viewing, and widescreen viewing is not where most reaction discovery happens anymore. Post a horizontal split screen to Shorts and you get two postage stamps floating in a black void.
Use it when you are making longer-form YouTube content that people watch on a laptop, and when your commentary is substantial enough that your face deserves equal billing.
3. Stacked vertical: footage on top, face below
Rotate the split ninety degrees and something clicks. On a 9:16 canvas, the trailer sits in the top portion at its native aspect ratio, uncropped, and your face fills the space underneath. Nothing is squeezed. Nothing is hidden. Both elements are large because vertical screens have height to spend.
This is the layout that survives cross-posting. The same file works on TikTok, Reels, and Shorts without a re-edit, and it still reads fine when someone watches it on a desktop feed. If you are choosing one layout to standardize on in 2026, this is it, which is exactly why the Reactr recorder outputs this format by default.
How big should your face be?
The rule of thumb that has held up: your face should be large enough that a viewer can read your expression without leaning in. On a phone held at arm's length, that means your head occupies roughly a third of the vertical space, not a tenth.
Genre pushes this around more than people expect. Horror trailers live on the delayed flinch, and the flinch is the content. When you sit down with something like Evil Dead Burn, your face needs room, because the moment your audience came for is happening on it, not on screen. Comedy behaves the same way: the laugh is the payoff.
Spectacle inverts it. A trailer for Spider-Man: Brand New Day is selling scale and choreography, and cropping that footage to make your face bigger actively costs the viewer something. Family animation sits in between. A Toy Story 5 teaser rewards a layout where both the frame and your reaction are legible, because the appeal is shared nostalgia, not one big shot.
Safe zones: the part everyone forgets
Every vertical platform overlays its own interface on your video. Captions, usernames, follow buttons, music credits, and a column of icons on the right side. If you place something important there, the platform will cover it.
Practical margins to work with on a 9:16 clip:
- Bottom 15 percent: reserved for captions and account handles. Keep your chin out of it.
- Right 15 percent: the icon column. Never put your face bubble in the bottom right corner of a vertical clip, which is exactly where the picture in picture habit puts it.
- Top 10 percent: assorted headers and search bars depending on the app.
The stacked layout dodges most of this by design, because the seam between the two panels sits in the middle of the frame where no platform puts anything.
Do not switch layouts mid-video
A tempting idea: start full screen on the footage, then cut to a big face for the reveal. It works in edited video essays. It rarely works in reactions, because the viewer loses track of where to look at exactly the moment you want their attention locked.
The version that does work is the pause. Hold the layout steady, stop the footage, and talk. The frame stays predictable while the pacing does the work. That is the entire reason pausing to comment is a native part of the Reactr recorder rather than something you fake in an editor afterward.
The one-minute decision
If you want this as a flowchart:
- Posting mainly to TikTok, Reels, or Shorts? Use stacked vertical. Stop reading.
- Making a 20-minute YouTube reaction people watch on a laptop? Split screen, with your panel on the left.
- Reacting to something where the imagery is the whole product and your commentary is sparse? Picture in picture, top left corner, and make the bubble bigger than feels natural.
- Not sure? Stacked vertical. It is the least wrong option in the most places, and it downscales into feeds without becoming unreadable.
Common questions
Is split screen better than picture in picture for reaction videos? On desktop, usually yes, because nothing is obscured. On mobile, no, because both panels end up too small to read. The aspect ratio of the destination decides this, not the layout itself.
What size should the reaction bubble be? Between 25 and 35 percent of frame width for a corner bubble. Smaller than that and expressions stop reading on a phone.
Do I need editing software to build these layouts? No. The Reactr recorder composes the vertical stacked layout while you record, in the browser, and hands you the finished clip with your handle on it. There is nothing to align by hand and nothing to render overnight.
Layout is a decision you make once and then stop making. Pick the format that fits where your audience actually watches, keep it consistent so your videos are recognizable in a feed, and spend the attention you save on the part that no template can do for you: having something worth saying when the trailer stops. Browse what is dropping right now, pick a trailer you have an opinion about, and record the first one tonight.