A study of 18,000 YouTube channels found Shorts help some niches and hurt others.

AIR Media-Tech tracked 18,000 English-language YouTube channels from March 2025 to March 2026, split between channels running long-form only and channels mixing in Shorts. There is no universal answer to whether Shorts help a channel grow. Fitness, food, business and education channels running a modest Shorts ratio, roughly 25 to 40 per cent of uploads, showed up to 41 per cent more subscribers than long-form-only channels in the same niche. Gaming, science, beauty and entertainment channels showed the opposite. Long-form-only channels held as much as 85 to 90 per cent more subscribers at scale than channels mixing formats.

Why this matters now

The Shorts debate has mostly run on anecdote. One creator adds Shorts and watches long-form views climb, credits the format, and posts about it. Another adds Shorts and watches the opposite happen, blames the format, and posts the reverse. Both are describing something real about their own channel, and neither is telling anyone much about a different one, because one channel's experience says nothing reliable about a different niche or a different audience.

What has been missing is scale, a sample wide enough that the outliers cancel out and the actual pattern is what is left.

What actually splits the winners from the losers

Niche decides it more than effort or execution alone. Fitness channels adding Shorts saw the strongest lift in the study, food and cooking close behind, and business and education channels benefited at a similar rate once Shorts sat in that 25 to 40 per cent band of total uploads. Push past 55 per cent and the pattern reverses even in a niche where mixing helps, Shorts start crowding out the long-form content the audience actually subscribed for. Gaming showed the widest gap in the whole dataset, with long-form-only channels carrying close to double the subscribers of channels mixing in Shorts at the same size. Two niches sat almost dead centre: travel, and home and DIY. For those, format mix moved next to nothing.

Where this leaves a channel working out its own strategy

Travel and home and furniture retail are two of the categories this data says do not move on format mix. Under one per cent of a subscriber difference either way is not a lever worth spending production time on. The lever that does move those channels is the one it always was: whether the metadata gives YouTube's classifier a clean, keyword-first signal of what a video is and who it is for, and whether the description and chapters back that signal up.

A golf or leisure club channel sits closer to the niches where the data shows a real gain than to the ones where it does not, closer to education and lifestyle content than to gaming or beauty. Worth testing a modest Shorts ratio there rather than assuming either way.

The bigger point sits underneath all of it. A format decision made on one viral video or one bad quarter is a decision made on noise. The channels that get this right test their own upload mix against their own niche's numbers, rather than copying whichever creator posted about it most recently.