Hook Formula: How to Start Your Video in the First 3 Seconds
Scroll TikTok for ten minutes and pay attention to what makes you stop.
It's not production quality. Half the videos that stop your thumb look like they were filmed in a bathroom. It's not the caption. You haven't read it yet. It's something in the first two seconds that creates a gap between what you're seeing and something you need to resolve.
That gap is the hook. And it's not magic — it's a handful of patterns that keep working because they're wired into how attention actually works.
Why three seconds and not five, or ten
Platforms optimise for watch time as a percentage of the video. A three-second watch on a ten-second video is 30% retention. A three-second watch on a sixty-second video is 5%. The algorithm treats these very differently.
This means the first three seconds aren't just about getting someone to watch your video. They're about signalling to the platform that your video is worth showing to more people. You're not just convincing the viewer. You're convincing the algorithm at the same time.
Three seconds is also roughly the time it takes for someone's brain to make a keep-or-skip decision. It's not arbitrary — it's the natural length of a single unit of attention.
The five patterns that actually work
1. The unfinished thing
Start mid-action. Mid-sentence. Mid-movement.
Not "today I'm going to show you how to sync cuts to the beat" — cut that entirely and open on the moment where a cut lands on a beat and the shot changes and something in your chest goes oh.
The brain hates incompleteness. If something is already in motion when you start watching, you stay to see how it resolves.
This is why the best TikTok hooks often start with "—and that's when everything went wrong" or show someone mid-fall before cutting to the beginning. You've already started a pattern. Stopping would leave it unfinished.
For music videos and creative content specifically: don't open on the artist standing still. Open on the moment. The turn, the impact, the peak of a movement. Let the viewer arrive in the middle of something happening.
2. The specific strange detail
Not "I'm going to show you something cool."
"The thing no one tells you about cutting on the beat is that you actually have to cut two frames before it."
Specificity does two things: it signals that you actually know something, and it creates a small open loop the viewer needs to close. Two frames before? Why? They have to stay to find out.
Vague hooks feel like clickbait because they are. The viewer's brain has been trained to distrust them. Specific hooks feel like something worth listening to.
The more counterintuitive the specific detail, the better. If it confirms what someone already thinks, they don't need to watch. If it contradicts something they assumed was true, they can't leave without knowing why.
3. The cost of not watching
"If you're still putting your best shots at the beginning of your music video, you're wasting them."
This hook works because it creates mild anxiety. There's something the viewer might be doing wrong. They need to know if it applies to them.
This isn't the same as clickbait like "you won't believe what happens next." That's vague. This is specific — it identifies a real behaviour (putting best shots at the beginning), names the consequence (wasting them), and leaves the explanation just out of reach.
The viewer stays because they need to either confirm they're safe or find out what to change. Either outcome is worth three minutes.
4. The visual that doesn't make sense yet
This one needs no words at all.
Show something that's visually interesting or unusual without explaining it. A shot that looks almost too good for the setting it's in. A cut that happens in a way you haven't seen before. A colour grade that makes you stop and wonder how.
The brain wants to explain what it's seeing. If the image doesn't come with an explanation, the explanation has to come from the video. So you watch.
This is why "process reveal" videos work so well — watching someone create something, with no context given upfront, is inherently unresolvable until the end.
5. The direct address that sounds like it's only for you
"If your edits feel like they're almost there but not quite—"
This hook works because it sounds like the person knows something about you specifically. It names a feeling — not a situation, a feeling — that a particular kind of viewer recognises from the inside. Anyone who doesn't have that feeling will keep scrolling. Anyone who does feel it stops, because someone just named their thing.
The mistake most people make with this pattern: they make it too broad. "If you're a video editor—" is not a hook. That's a demographic. "If your edits always feel one decision away from being good" is a feeling. The more specific the feeling, the more it resonates with the exact person you're talking to.
What kills a hook before it starts
Starting with your name and who you are. Nobody cares yet. That's not rudeness — it's just that you haven't earned that attention. Say who you are after you've given someone a reason to want to know.
A slow zoom or fade in. You have three seconds. Spending one of them on a stylistic transition is a bad trade. Cut to the first frame clean.
A long setup sentence. "So today what I wanted to talk about was the way that edits can sometimes feel a little bit off and you're not sure why..." is nine seconds of ground you could have covered in two words.
Asking a question that doesn't cost anything to ignore. "Have you ever wondered about X?" is only a hook if not knowing the answer creates real discomfort. "Have you ever wondered about the history of the sandwich?" does not create real discomfort.
One thing worth knowing about hooks and honesty
The hook makes a promise. The rest of the video has to keep it.
A hook that's stronger than the content creates a specific kind of disappointment — the viewer feels manipulated, and they leave that feeling with your channel, not just the video. On short-form platforms that feeling spreads fast.
The best hooks are ones where the hook is the content — a genuine surprising thing, genuinely stated, with the actual explanation following. Those videos hold watch time, not just the first three seconds. And watch time is what actually moves the algorithm.
Making it practical
Next time you sit down to edit, write the first sentence of your video down before you cut anything. Then ask: if someone read only this sentence and nothing else, would they have a reason to read the next one?
If the answer is no — and it usually is on the first draft — cut the sentence and try again. The hook is almost never the first thing you write. It's usually the fourth or fifth.