B-roll for UGC ads: the cutaway is where the product sells

The talking head is the A-roll. It is not the ad.

In UGC, B-roll means everything that is not the creator talking to camera: hands using the product, the unboxing, the texture close-up, the lifestyle moment that shows where the product lives. It solves two problems at once. It keeps the viewer watching, because the frame keeps changing, and it makes sure the product is actually on screen for more than a glance.

Watch a batch of underperforming UGC ads and the pattern is nearly universal: forty seconds of a face, one product shot, a link. The creator was good. The ad never showed the thing. Viewers do not buy a face; they buy what the face is pointing at, and if the camera never points there, the ad is a monologue.

This matters even more with AI creators. A synthetic performer holding the frame for thirty seconds gives the viewer thirty seconds to notice something is off. Cut away every few seconds and the performer becomes what it should be: a narrator over proof.

How often to cut away

The working rule on TikTok and Reels is a scene change every two to three seconds, which means a thirty-second UGC ad usually cuts to B-roll six to ten times, each for a second or two, while the creator's audio keeps running underneath. The audio carries the argument. The picture carries the proof.

• Under the hook: rarely. Let the face land the first line, then cut.
• Under the problem: a moment that shows the pain, not the product.
• Under the claim: the product doing the thing being claimed, in real time.
• Under the proof: hands, results, comparisons, the screen if it is software.
• Under the close: the product at rest, packaging, the moment of use that the viewer will imagine.

Shots worth keeping in a library

The most reusable B-roll is not product-specific. It is the connective tissue: a reaction, a hand reaching for something, a scroll on a phone, a doorstep delivery, someone nodding along, a walk through a kitchen. Those clips sit under almost any line. Tag them by function and emotion rather than by object, because the question during an edit is never "do I have a clip of a mug", it is "do I have a calm, satisfied moment for this line."

A working library for one brand is smaller than people expect: thirty to fifty clips covering those six functions in two or three energy levels will carry months of ads. What matters is that every clip is vertical, looks phone-shot, and features people who could plausibly be in the same room as your creator.

Product footage you should shoot yourself

Library footage covers the connective tissue. It cannot cover your product. Plan one afternoon with a phone and good window light and shoot the product moments you will reuse in every ad for a year.

• Unboxing, twice: once fast, once slow.
• The product in hand, turned slowly, at arm's length and close.
• The product in use, in real time, from the user's point of view and from across the room.
• The result, whatever that means for you: a clean surface, a screen, a before and after, a full plate.
• The product at rest in three settings.
• Packaging details: label, ingredients, size against a hand.

Shoot vertical, at the same frame rate you edit in, and without music. Ten minutes of this footage, tagged by function, is worth more than any stock subscription.

Making stock stop looking like stock

Stock footage fails in UGC for three reasons: it is horizontal and cropped, it is lit like a commercial, and everyone in it is too happy. Feed-native B-roll is vertical from the start, shot on something that looks like a phone, and populated by people who behave like the creator in the A-roll. That last point is the one people miss. A cutaway with a different energy than the talking head reads as an edit. A cutaway with the same energy reads as the same story.

This is why the VibesUGC library is organized by creator, emotion, setting, and delivery rather than by keyword. You are usually looking for a moment that matches a line and a mood, and the footage is vertical and social-native by design. Browsing and selecting clips is free; a paid render starts only when you explicitly launch one.

• Match the color temperature. Warm A-roll with cool B-roll is the fastest way to look assembled.
• Match the camera motion. Handheld talking head, handheld cutaways. Locked-off talking head, locked-off cutaways.
• Keep cutaways short. Two seconds of a slightly wrong clip is fine; six seconds is a different ad.
• Never let a stock person look at the camera. That is the creator's job.

Plan the cutaways before the script

The order most teams use is script, shoot, edit, then hunt for B-roll. Reverse it. Decide the beats, pull the cutaways that prove each beat, and then write the lines that sit over them. The script gets shorter, the product gets more screen time, and the edit assembles itself.

A practical way to do this: lay the cutaways on a timeline first with placeholder gaps for the creator, then write one line per gap. Each line has to do one thing: set up the next cutaway. You will find you cannot write a long line, because there is no room for one, and that constraint is exactly what makes the ad feel native.