Choosing B-Roll That Earns Its Place

Open almost any explainer video and count the cutaways that show exactly what the narrator is saying. Someone mentions a team, and there is a shot of people at a table. Someone mentions growth, and there is a line going up. Nothing is wrong with any individual shot, and together they have added nothing at all — the viewer had the information from the sentence, and the picture just spent three seconds agreeing with it.

B-roll is expensive. It costs shooting time, storage, edit time, and a slice of the viewer's attention that the main footage no longer gets. A shot has to return more than that, and there are only three ways it can. This guide covers what those three jobs are, why the most common one is also the weakest, and a test that rules out most of the footage you were about to use.

Three Jobs, Not One

Nearly all amateur B-roll is illustration, and illustration has the narrowest window of the three. It works when the subject is genuinely unfamiliar — a machine the viewer has never seen, a process with steps, a place with a specific character. It stops working the moment the viewer could have imagined the thing themselves, which for most abstract nouns is instantly.

Illustrate Only What Cannot Be Imagined

The useful test is not "is this relevant" but "does this show something the sentence could not deliver".

A shot of a specific workshop with its own clutter and its own light shows something no sentence carries. A shot of a generic office while the narrator says the word office carries nothing. The difference is specificity, and specificity is why footage shot for the project beats footage found for it.

There is a practical consequence for what you shoot. Wide establishing shots are the least useful thing on most B-roll cards, because they show the category rather than the instance. Close detail is the most useful: hands doing the actual task, the surface of the actual object, the particular mess of the actual room. Detail is what a viewer cannot supply from imagination.

Contrast Does the Most Work

The strongest cutaways sit at an angle to the narration rather than on top of it.

Someone describes a process as simple, and the picture shows the eleven steps it actually takes. Someone talks about the finished product while the shot stays on the version that failed. The narration says we tested it, and the picture shows the tenth test, at night, with the room a mess. None of this contradicts the speaker; it adds the texture the sentence had to leave out.

This works because the viewer is doing something. Illustration asks nothing of them. Contrast asks them to hold two things at once and resolve them, and that small act of resolution is attention you did not have to buy any other way.

The line to watch is tone. A contrast that undercuts the speaker reads as sarcasm, and unless you are making that argument deliberately it turns the audience against the video rather than into it. The distinction is whether the picture adds a complication the speaker would recognise, or one they would object to.

Conceal, and Be Honest About It

The third job is the one that gets the most use and the least discussion: covering something.

A cutaway hides a jump cut where two takes were joined. It bridges a removed stumble so the speech does not visibly skip. It covers the moment where the camera was reframed, or the lighting changed, or a question was cut out. Cleaning up speech nearly always creates one of these seams, and the shot that covers it is doing real structural work — the process that produces them is described in removing filler words without making speech sound chopped.

Naming the job matters because it changes the requirement. A concealing shot has to be plausible at that instant and it has to last exactly as long as the seam. It does not have to be beautiful and it does not have to be thematically apt. Editors who forget this spend an hour looking for the perfect shot to cover half a second, when almost any related detail would have worked.

The Test

Before a cutaway goes in, finish this sentence out loud: this shot is here because ————, which the main footage could not do.

If the sentence will not finish, the shot is decoration. If it finishes with "it looks good", the shot is decoration with better lighting. The test takes two seconds and removes a surprising proportion of a B-roll bin.

Three follow-up checks catch the rest.

Does it survive being on screen? Cut it in and watch at normal speed without scrubbing. Shots that seemed fine in the browser often turn out to be soft, shaky, or over in half the time you expected.

Does it match? Footage from a second camera, or from a different day, arrives at a different exposure and colour temperature, and the mismatch reads as a mistake rather than as a change of angle. Correct it before you judge the shot — the order for that work is in colour correction versus grading.

Does it hold the pace? A cutaway is a shot like any other and obeys the same length discipline as the rest of the sequence. Dropping a leisurely eight-second insert into a section cut at two seconds stops the video dead, which is a pacing question covered in pacing a five-minute video.

Shoot For It, Not Just Around It

Most B-roll problems in the edit are shooting problems, and they are cheap to fix at the source.

Shoot detail before wides, because detail is what you will actually use. Get every insert twice, once static and once with a slow move, so you have a choice when the neighbouring shots turn out to be busy or still. Hold each take far longer than feels necessary — a shot you can only use for two seconds is a shot with no trim room, and trim room is what lets you place a cutaway on a specific word.

Record ambient sound while you are there, even when the shot will be used mute. Room tone from the actual location is the cheapest way to make a concealing cutaway stop feeling like a patch, and it takes thirty seconds to capture.

Finally, shoot the failures. The rejected version, the messy bench, the take that went wrong. That is where contrast footage comes from, and it is the material nobody thinks to keep.

Next Steps

Open your last finished video and run the test on every cutaway in it. Say the sentence out loud for each one. Count how many finish, and count how many were illustration of something the viewer could already picture.

Then look at what your B-roll bin is mostly made of. If it is mostly wide establishing shots, the next shoot should be almost entirely close detail — that single change does more for the edit than any amount of selection discipline afterwards. How those shots are joined is a separate skill, covered in cutting on action, and the surrounding craft sits under editing craft and workflow.