He heard somebody say that a video over ten minutes performs better. His subject needed five minutes, so he stretched it: he added an introduction about himself, re-explained the idea two ways, and told an unrelated story.
He published it, and the retention curve fell off a cliff at minute four. In the comments, one sentence kept repeating: "the useful part starts at minute seven."
Then on another video he did the opposite: a subject needing twelve minutes, cut to six. It came out incomplete, and the comments asked about the things he had removed.
Length is not a number you aim for. Length is a result.
The rule: length follows the number of stations
Before writing, list the video's stations: the things the viewer has to understand, in order. Each station needs a minute or two depending on its difficulty.
Then add them up: five medium stations plus an opening and an ending comes to seven or eight minutes. That is your video's length, not the number you heard from somebody.
If the count comes out very large, the subject is wider than one video. If it comes out very small, the idea is still a subject rather than an idea.
How to estimate length from your script
Read a page of your script out loud at your natural pace and time it. Now you know how many minutes a page is for you, rather than for somebody else.
Then use that number on every video, and add thirty percent if the video has shots without speech or location scenes. Most people are surprised to find their script is longer than they thought.
Four things that set the length before any rule
| The factor | Raises the length | Lowers it |
|---|---|---|
| Kind of video | A film, a story, a complex explanation | Answering a question, one piece of advice |
| The viewer's state | Searching in order to learn | Scrolling with no purpose |
| The platform | A long-form channel | Short clips |
| Your material | You have scenes and interviews | You have talking to camera |
The second factor matters most. Somebody searching for an answer accepts twenty minutes if the answer is complete; somebody scrolling leaves at second twenty however good your video is.
Neither stretch nor cut
What stretches a video with no value: an introduction about yourself, re-explaining what was understood, an unrelated story, repeating the idea two ways, and a long thank-you at the end.
What must not be cut: a station the understanding needs, an example that fixes the idea, the source of a fact, and a caution that belongs to a sensitive subject.
The test is one: if you delete something and the understanding does not suffer, it was surplus. If it suffers, it was structural.
The first minute decides more than the length
A viewer does not decide by the duration shown but by the first minute. If they feel in that minute that you know where you are going, they will watch ten. If they feel you are hesitating, they leave even if your video is three minutes long.
So investing in the first minute matters more than any decision about length, and the title and description are what decide whether the video is opened at all, as in Writing the Title and Description.
Read the right length out of your own numbers
Do not ask the internet for the ideal duration, because you have a more accurate answer: open your last ten videos and write a two-column table: duration, and percentage viewed.
A pattern will appear: a length past which the percentage drops sharply. That is your length, with your audience, on your subject. How to read the retention curve is in The Retention Curve in YouTube Analytics.
And beware of confusing two numbers: watch time and watch percentage are different things, and shortening a video raises one without improving anything, as explained in Watch Time and Watch Percentage.
When to split the subject in two
- When it contains two independent parts, each understandable on its own.
- When your stations pass eight and each needs two minutes.
- When the second half serves a narrower audience than the first.
Do not split a video to raise your count, because a second part that does not stand alone gets few viewers. And a video ending with "in the next part" needs that next part to actually exist and to arrive soon.
How those parts sit in your monthly plan is in A publishing calendar that holds.
And in short clips
The rule inverts: there, duration is a ceiling rather than a result. Start from the limit you want to stay under, then pick an idea that fits it. An idea that cannot be completed in that time is not cut down; it is left for the long video.
Three minutes before exporting
Ask three questions of the final cut:
- Which minute can I delete without losing anything? Delete it.
- Where would a viewer leave if this were somebody else's video? Fix that spot.
- Does the title's promise land before the middle? If it lands at the end, you are testing people's patience.
After publishing, record the duration in your notes along with what happened, and repeat that reading monthly as in After Publishing: Reading a Video's Performance.
The rule in one line: make your video as long as the subject needs, not as long as somebody told you.