B-roll: the supporting footage that saves retention, and where the name comes from

In the film era, an edit used two reels running in parallel. The A roll held the main scene, usually someone talking. The B roll held the complementary material: landscapes, details, supporting shots.

The name outlived the technology. Nobody uses reels today, but the distinction is still exact: A-roll carries the narrative, B-roll illustrates it.

In social content the A-roll is usually a talking head, and the B-roll is everything layered over it.

The reason is physiological before it is aesthetic. A static image with no change makes attention drop fast, and in short video that is fatal.

B-roll solves it three ways:

• It renews visual stimulus. Every image change restarts attention.
• It hides the splices. A cut in the speech is obvious on A-roll and invisible when covered by B-roll. That is what lets you record in pieces and cut the rambling without the video looking chopped.
• It proves what is being said. Showing beats claiming, and an image of the result is worth more than a description of it.

In practice it is the highest return per effort item in Reels editing.

The most common question is what exactly to shoot. Four categories cover almost everything:

• The action itself. Your hands doing what you are describing. It is the easiest B-roll and the most convincing.
• The detail. A close shot of the object, the screen, the product.
• The context. The environment, the route, the before and after.
• Screen recording. In digital content, recording your screen is B-roll and almost nobody treats it that way.

A practical rule on quantity: shoot three to five seconds per take and collect more takes than you think you will use. Surplus is cheap; a gap means filming again.

• Generic stock B-roll. The stock shot of people smiling in an office illustrates nothing and communicates laziness.
• Changing images too fast. A cut every half second tires viewers and prevents them understanding what is being shown.
• B-roll that contradicts the speech. If the image shows one thing and the audio says another, people watch the image and lose the argument.
• Forgetting the audio. Cutting the sound along with the image produces a noticeable jump. Keeping the speech continuous underneath fixes it.

• I shot more takes than I need.
• Each take runs three to five seconds.
• The B-roll illustrates what is being said rather than contradicting it.
• I use B-roll to hide the splices in the speech.
• I do not use generic stock footage.