How long should a YouTube Short be?

The hard limit, and why it is not the answer

At the time of writing, YouTube's help centre says that any video uploaded on or after 15 October 2024 with a square or vertical aspect ratio and a length of up to three minutes is categorised as a Short, and that these videos are monetised through the Shorts revenue sharing model. Before that date the cap was sixty seconds, which is why so much advice online still says "under a minute". It is out of date.

Two details in that rule matter more than the number. First, YouTube decides what is a Short by shape and length, not by which upload button you pressed. A vertical video that runs three minutes and one second is not a Short; it is a tall long-form video, judged by long-form standards. The same page says that if you do not want a video treated as a Short, the way out is a wider aspect ratio such as 16:9, not a different upload route. Second, the cap is a ceiling, not a target. Nothing in the rule rewards you for getting close to it.

Music is where the limit bites in practice. At the time of writing, YouTube's music eligibility page for Shorts says that Shorts longer than one minute with an active copyright claim will be blocked, regardless of the claim's policy. On a sub-minute Short a claimed track usually just shares revenue; on a ninety-second story it stops the video being seen at all. The same page says that, depending on the song, you may be able to add up to ninety seconds of music from the Shorts audio library to a three-minute Short, and that the picker shows how many seconds a track will cover. Royalty-free music from the YouTube Audio Library is the alternative it offers for anything longer.

What the feed measures instead of length

The Shorts feed does not have a preferred duration. It has a preferred behaviour: people keep watching, and when the video ends they do not swipe away annoyed. A twenty-second video that is watched to the end and replayed will be shown to more people than a two-minute video most viewers abandon at forty seconds, even though the second one collected more watch time per viewer.

This is why the public view count tells you very little now. At the time of writing, YouTube's help centre says that since 31 March 2025 a Shorts view is counted when a Short starts to play or replay, with no minimum watch time, and that Partner Program eligibility and Shorts revenue sharing are based on "engaged Shorts views" instead. The stricter figure is still in YouTube Analytics under the name engaged views. When you are judging whether a length works, engaged views and the retention curve are the numbers to read, not the one under the thumbnail.

There is a floor as well as a ceiling. A video that is over before the viewer has worked out what it is about gives the feed nothing to measure and the viewer nothing to react to. Very short Shorts can do well, but only when the idea is complete: a single fact with a punchline, a one-line reveal, a before-and-after. Shortness is not a hook by itself.

A working range for each faceless format

These ranges are not rules; they are where the formats tend to settle once everything that does not move the video forward has been cut. Treat them as the starting point for your own testing, which the last section covers.

The pattern underneath all of them is the same: length is set by how many beats the idea has. A fact has one beat. A story has a setup, a turn and a payoff. A ranked list has as many beats as items, which is why lists drift long and are usually better with fewer, stronger entries.

• Single fact or "did you know" (animal facts, space, science): roughly fifteen to thirty-five seconds. One claim, one visual, one closing line. If you have three facts, make three Shorts.
• Pick-one, would-you-rather and quiz formats: fifteen to thirty seconds. The viewer's answer is the ending; do not pad after it.
• Short explainers (psychology, finance, stoic ideas): forty-five to ninety seconds. Room for a claim, one concrete example and a takeaway; not room for a second example.
• Single-story formats (Reddit stories, text-message stories, scary stories): sixty seconds to about two minutes. The length is set by the story's turn; if the twist lands at seventy seconds, end at seventy-five.
• True crime, history and mythology narratives: ninety seconds to the full three minutes, but only when every beat is doing work. These are the formats where the three-minute cap is genuinely useful, and the ones most tempted to waste it.

Let the script set the length, not the other way round

The most common mistake is deciding on a duration first and writing to fill it. The result is a script with a slow second act, because the writer needed another forty seconds. The better order is to write the tightest version of the idea, time it, and accept whatever length that produces.

Timing a script is arithmetic. A synthetic or human voice at a natural narration pace lands somewhere around two and a half words a second, a touch faster for energetic styles and slower for horror or calm formats. So seventy-five words is about thirty seconds, a hundred and fifty words is about a minute, and a three-minute Short needs around four hundred and fifty words of actual narration. If your draft is six hundred words and you want it to stay a Short, the fix is cutting, not speeding the voice up until it sounds like a disclaimer.

This is also how a tool-made video gets its length. In FacelessVideoCreator, the duration of the finished video is a consequence of the script's word count and the voice's pace; there is no separate dial that stretches or compresses the result. If a video comes out longer than you wanted, the answer is a shorter script, and that is true whether a person or a model wrote it.

A practical cutting rule: read each sentence and ask what the viewer would lose if it vanished. "In this video we're going to look at" loses nothing. "But that is not what the police found" loses the story. Keep the second kind. Then check the pauses: a horror script with a dramatic beat of silence after every line can gain twenty seconds of nothing, so time the finished voiceover, not just the words.

The first three seconds and the last three

Length is judged at two points: the moment a viewer decides whether to stay, and the moment they decide whether they were glad they did. Everything between those points matters less than beginners think, and both ends matter more.

At the start, the viewer has not committed to anything. A title card, a channel logo, or a sentence of throat-clearing is where most swipes happen, and those early swipes drag down every other metric regardless of how good the rest is. Open on the claim, the question or the first line of the story. If the video is a two-minute narrative, the first line should make the length feel like a promise rather than a cost: "she had three minutes to decide" earns a long video; "so today's story is about" does not.

At the end, stop at the payoff. A closing ten seconds of "like and subscribe for part two" is a sixth of a minute-long video spent on something the viewer did not come for, and the retention graph will show it as a cliff. If you want a call to action, make it one line and put it after the last beat, not before it. A Short that ends cleanly also loops back to its own opening without a jolt, and a replay counts in your favour.

If you post the same video elsewhere

Most faceless creators post one file to several feeds, and each has its own cap. At the time of writing, TikTok's help centre says videos recorded in the TikTok app can be up to ten minutes and videos uploaded to TikTok can be up to sixty minutes, so a three-minute Short fits easily. Instagram's help centre says, at the time of writing, that a Reel can be up to twenty minutes long, but adds a note that Reels over three minutes will not be recommended to new audiences. That note is worth more than the cap: it means the three-minute line is effectively the same on both YouTube and Instagram if you want to reach people who do not already follow you.

The limit is rarely the problem. The problem is that the same length does not perform the same way everywhere. A two-minute narrative that holds on YouTube, where viewers are used to longer videos, can struggle on feeds built around shorter clips. If you auto-post to YouTube Shorts and TikTok from one upload, a sensible default is to keep videos under about ninety seconds until you have evidence that longer ones hold on each feed. That ninety-second line also happens to be the most you can usually cover with Shorts audio library music, so it is a convenient place to draw it. Longer story formats can go to YouTube alone, or be cut into parts for the others.

If you do split a story into parts, each part needs its own hook and its own small payoff. A part two that only makes sense after part one will be watched by the small fraction of people who happened to see both, and that is not how feeds work.

Read more: How to make faceless TikTok videos

Find your own number from your analytics

Whatever you read here or elsewhere, your audience will have its own tolerance, and the only way to learn it is to test. The test is cheap: pick one format, make a batch at one length and a batch at another, and compare them after a couple of weeks using the same metrics for both.

Compare engaged views rather than views, because views now count any play. Compare the share of viewers still watching at the end, which the retention graph shows directly. Compare how many swiped away in the first few seconds, which the Shorts analytics in YouTube Studio report separately and which tells you about the hook rather than the length. Then look at where the retention curve drops: if the forty-second version holds to the end and the ninety-second version collapses at fifty seconds, the audience has told you the answer, and it is not "make them longer".

Run the test with enough videos that one lucky hit does not decide it. Ten of each is a reasonable minimum. Change one thing at a time: a batch that is both longer and uses a different voice tells you nothing about either. Once you have a number that works, it becomes your default, and you revisit it when the format changes, not every week.

Questions

Does a longer Short earn more money?

Not automatically. Shorts revenue is based on how the video performs in the feed, not on how much footage you uploaded, and at the time of writing YouTube says that performance is measured in engaged Shorts views rather than raw plays. A long video that most viewers abandon early is shown to fewer people, which usually means less revenue than a short one watched to the end. Length within the three-minute cap does not change eligibility either way.

Is a 16:9 video under a minute a Short?

No. At the time of writing, YouTube's rule is about shape as well as length: a Short is square or vertical and up to three minutes. A widescreen video of any length is treated as ordinary long-form, and YouTube's own advice for keeping a video out of the Shorts feed is to use a wider aspect ratio such as 16:9. If you want a clip in the Shorts feed, it has to be vertical or square, however short it is.

I have no analytics yet. What length should I start with?

Start in the middle of the range for your format rather than at either end. For most faceless formats that is forty-five to sixty seconds: long enough that the retention curve shows you where people leave, short enough that you can make ten of them quickly and have something to compare. Once you have a dozen videos, look at where the curve drops and adjust from there.

Do burned-in captions change how long a Short should be?

Not the total length, but they set a floor on pace. Many viewers watch with the sound off and read the captions, and text that changes faster than roughly three words a second becomes hard to follow. If a script is narrated quickly to squeeze it under a target, the captions will flash past and the silent viewers leave. The fix is the same as everywhere else in this article: cut words, not seconds.

Does a Short need to loop?

There is no rule that says so, but a clean loop helps. At the time of writing, YouTube counts a replay as a view, and a video that ends on its last beat and starts again on its hook is replayed more often than one that trails off through a call to action. You do not need a gimmick where the last frame matches the first; you need an ending that stops when the idea does.

Can I cut a Short out of my own longer video?

Yes, and for faceless channels that make both it is a cheap source of Shorts. The clip still has to be square or vertical and under three minutes to be treated as a Short, and it needs to stand alone: a moment that only makes sense with the previous ten minutes of context will not hold a viewer who has never seen the long video. Pick the single strongest beat, open on it, and end when it lands.

Sources

Platform rules change. These pages are the authority; this article summarises them as of the date above.

• YouTube Help: Understand three-minute YouTube Shorts
• YouTube Help: Music eligibility for YouTube Shorts
• YouTube Help: Get started creating YouTube Shorts (changes to how Shorts views are counted)
• TikTok Help Center: Camera tools (video length)
• Instagram Help Center: Record a reel on Instagram (how long your reel can be)