How long is a hook?

Somewhere near the start of most channels there is an intro. Five seconds of logo, a whoosh, maybe a strapline. It was made carefully, either by the owner in a template or by somebody who was paid for it, and it goes on the front of everything.

It is also, on a scrolling feed, the entire decision. The viewer has watched your branding and nothing else, and then they have gone.

What is the viewer deciding?

Not whether to watch your video. That is the mistake most intros make.

On a feed, the viewer is deciding whether to stop scrolling. They may not have chosen the subject, so the opening needs to make clear what is here and why it might interest them. Give them enough to recognise a useful lesson, a good story or an answer they were looking for.

On a platform where somebody has clicked a title, the decision is different again. They have already stopped. What they are now doing is checking whether the video is about to deliver the thing the title promised, which is a question you can answer badly for a while before they leave.

Two different decisions, two different windows.

How long is the window on a scrolling feed?

Take a second and a half as a working figure, and treat it as exactly that.

The one-and-a-half-second figure is a house prompt for reviewing feed openings, not a platform rule. Its purpose is to ask whether the subject becomes clear quickly; the actual duration depends on the idea and response.

Which gives you a straightforward test for any opening. Play it, count to two, and stop. Whatever happened in that time is the whole of your case.

Once somebody has clicked

A deliberate click gives you a promise to fulfil. There is no fixed number of seconds that suits every opening.

Somebody who has clicked a title and a thumbnail arrives holding a promise. The opening's job is to confirm that the promise is about to be kept and to make the shape of the video clear. That takes more than two seconds and it can be done comfortably in fifteen. What loses them at this stage is not slowness exactly. It is the suspicion that the thing they came for is a long way off.

So the same principle applies at a different scale. Start on the subject, not on yourself.

What does an intro animation cost?

The whole window, on the fast platforms, and the most valuable part of it everywhere else.

There is a version of this that sounds harsh and is just arithmetic. If the window is roughly the length of a doorbell, and the first thing in the video is a five second logo sting, then the viewer's entire experience of your work is a logo. It does not matter how good the logo is. It is answering a question nobody asked, at the only moment they were asking a different one.

The window gets spent on branding because branding feels like the professional thing to lead with, and because the alternative feels abrupt. It is abrupt. That is the point.

Where should the branding go instead?

Later, smaller, or nowhere at all.

A recognisable voice, colours or presenter may already establish the brand. A logo sting earns its place when it adds to the opening; it can move later or come out when it delays the subject.

Options include a small mark, a brief bumper after the opening promise or an end card. Check that each serves the piece: viewers can still be deciding what to watch or do at the end.

What to start on instead

The thing they came for. Three shapes work, and all three can be filmed in a single take:

1. The result, shown before it is explained. The finished dish, the repaired joint, the before and after, on screen while you say what it is. 2. The question, asked in the viewer's own words. Not the technical name for the problem, the way they would describe it to a friend. 3. The objection, said out loud before they say it. Naming the reason somebody would scroll past is a surprisingly reliable way of stopping them.

Go and look at the first two seconds of your last upload. Not the first ten. The first two, because on a scrolling feed those opening seconds are often all you get. If those two seconds do not contain the subject of the video, you already know what to change, and the change costs nothing but a re-edit.

Which of those three openings works for your audience is the next question, and it is answered by filming all three and reading what held.

Sources and evidence

What here is a platform fact and what is our own method. Platform behaviour was checked on 4 September 2026.

Standards and platform facts. YouTube's own analytics documentation gives retention and reach metrics defined meanings, so a creator can see, second by second, roughly where viewers left. That same documentation is careful to say the definitions describe viewer response; they do not by themselves explain why a change happened. That is the honest limit on the numbers in this piece. The second and a half, and the fifteen seconds, are not published anywhere as a platform specification, because no platform sets one. They are working figures this studio uses, arrived at by reading that kind of retention data across many openings, not a rule handed down by anybody.

Our method. Playing an opening back, counting to two, and asking what a stranger would have seen in that time, then filming more than one version and reading what the retention curve does with each, is how this studio treats the window. It is something to measure on your own channel, not a number to borrow from ours.

References.

• Bruns, D. et al. Measuring gaining and holding attention to social media ads with viewport logging: a validation study. tandfonline.com. Attention to social-feed ads can be extremely brief; viewport logging tracks gaining/holding attention. Supports 'feed attention is brief', not a specific 'most viewers gone by two seconds' figure.
• YouTube Help. Measure key moments for audience retention. support.google.com. How the retention report reads (flat/decline/spike/dip, first-30s intro, new vs returning). No universal benchmark; does not say a given edit tactic lifts retention.