How long should a YouTube video be?

Eight minutes is a real revenue threshold and a terrible content target. Here is what the length studies found, what YouTube's own rules say, and how to pick a number for your channel.

8 minute read

The short answer, and the reason it is unsatisfying

There is no single right length, and the honest version of the answer is that length is a consequence of your idea rather than a setting you pick beforehand. That said, two numbers do real work. Eight minutes is the threshold at which YouTube lets you run mid-roll ads, which is a genuine revenue cliff. Three minutes is the line where a vertical upload becomes a Short instead of a video. Everything between and beyond those is a judgement call about whether your material can hold attention for that long. The mistake almost everyone makes is inverting the order: choosing a duration first, then padding or amputating the content to hit it.

What the large studies actually found

Backlinko's analysis of 1.3 million YouTube videos found that the average length of a video ranking on the first page of YouTube search is 14 minutes and 50 seconds. A more recent dataset from AIR Media-Tech, covering 18,080 English-language channels and roughly 4.5 million videos published in 2025 and early 2026 with Shorts excluded, points the same direction: longer videos out-engage shorter ones in most niches, and the 5 to 10 minute band that most creators default to is not where the strongest engagement sits in any niche or at any channel size. It also found the peak differs sharply by category — gaming and DIY hold up around 20 to 30 minutes and longer, while business and finance content peaks at durations most creators would consider unpublishable.

Why those findings are weaker evidence than they look

Both datasets measure correlation, and the causal story runs at least partly the other way. Channels confident enough to publish 25-minute videos tend to be channels with material worth 25 minutes and an audience that already trusts them. A first-page result is long because comprehensive answers rank, not because YouTube rewards duration. If you take a seven-minute idea and stretch it to fifteen, you do not inherit the performance of the fifteen-minute videos in the study — you inherit the retention curve of a padded seven-minute video, which is worse than the original. Read these numbers as permission to go long when the content justifies it, not as a target to hit.

The 8-minute rule is real, and it is about money

This one is not a heuristic, it is written into YouTube's product. YouTube's help documentation states that on monetised videos of eight minutes or longer you can turn on mid-roll ads. Shorter videos are limited to pre-roll, post-roll and overlay formats. The threshold used to be ten minutes; Tubefilter reported YouTube lowering it to eight in July 2020, explicitly to reduce the incentive for creators to pad videos just to clear the bar. Whether it worked is debatable — a video at 7:59 still earns meaningfully less than the same video at 8:01, and that gap is why so much YouTube advice fixates on eight minutes. It is a legitimate consideration if you are monetised. It is irrelevant if you are not.

What changed about mid-rolls in 2025

The eight-minute threshold survived, but how the ads inside it are placed did not. From 12 May 2025, YouTube began prioritising mid-roll slots at natural breakpoints — pauses, transitions, scene changes — instead of dropping them mid-sentence or mid-action, and it retroactively added slots at natural breaks to older videos with manual placements, with an opt-out in Studio. YouTube said channels running automatic placements alongside manual ones saw around a 5 percent increase in ad revenue compared with manual-only. Search Engine Journal's coverage of the announcement is the clearest summary. The practical effect for length: a video with genuine structural breaks monetises better than one long unbroken monologue of the same duration.

Retention percentage versus total watch time

The two metrics pull in opposite directions and creators regularly optimise for the wrong one. A five-minute video held to 60 percent delivers three minutes of watch time. A twenty-minute video held to 35 percent delivers seven. YouTube's distribution system leans on total watch time and, as Buffer's guide to the algorithm sets out, on satisfaction signals layered on top of it — surveys, returns, whether the session continues afterwards — rather than on raw duration. So a longer video that holds a mediocre percentage often out-distributes a short one with a lovely retention graph. What you cannot do is fake it. Padding lowers the percentage without raising the total, which is the worst of both.

Shorts are a different question entirely

Anything three minutes or under and vertical is a Short, with its own feed, its own audience behaviour and a much lower ad payout per view than long-form. That makes the choice strategic rather than technical: Shorts are cheap reach, long-form is where watch time and mid-roll revenue live. The productive pattern is not choosing between them but sequencing them — publish the long video, then cut the two or three self-contained moments out of it that stand alone, and let those do discovery work while the full version does the earning. That way the length of the main video is set by the material, not by what will survive a vertical feed.

How to pick your number

Open YouTube Studio and sort your last thirty videos by average view duration in minutes rather than by percentage. You are looking for the band where your own audience reliably gives you the most minutes, which is usually narrower than you expect and rarely matches general advice. Then make length a per-video decision: cover the idea completely, cut everything that exists only to fill time, and check the runtime afterwards. If a genuinely complete video lands at 7:30 and you are monetised, look for real material to add rather than padding. If it lands at 25 minutes and retention holds, publish it. Once the long video exists, batching the vertical cuts out of it is a separate job — a clipping workflow like FrameOS handles that pass so the main upload never gets shaped around what makes a good Short.

FAQ

Is 8 minutes still the minimum for mid-roll ads on YouTube in 2026?

Yes. YouTube's help documentation still states that mid-roll ads can be enabled on monetised videos of eight minutes or longer, with shorter videos limited to pre-roll, post-roll and overlay formats. The threshold was lowered from ten minutes to eight in 2020 and has not changed since. What did change, in May 2025, is where YouTube places those mid-rolls.

What is the ideal length for a YouTube video?

There is no universal ideal. Backlinko's study of 1.3 million videos found first-page search results average 14 minutes 50 seconds, and larger recent datasets show longer videos out-engaging shorter ones in most niches, but both measure correlation. Treat the data as permission to go longer when your material supports it, and let the idea determine the runtime rather than the reverse.

Do longer YouTube videos get more views?

Longer videos tend to accumulate more total watch time, which is what YouTube's distribution leans on, so they often reach more people even at lower retention percentages. That is not the same as length causing views. A padded long video produces a worse retention curve without more watch time, so the advantage only holds when the extra minutes are genuinely worth watching.

Should I make Shorts or long-form videos?

Both, in sequence rather than instead of each other. Shorts are three minutes or less, vertical, and pay far less per view, but they reach new viewers cheaply. Long-form is where watch time and mid-roll revenue accumulate. Publishing the long video first and cutting Shorts out of it afterwards gives you both without letting the vertical format dictate the main upload.

Sources

Related pages