The first 3 seconds: how to write a video hook that stops the scroll

You shot the video yourself, edited it, showed the product at its best — and got thirty views, three of them from people you know. Most owners go looking for the problem in the edit, the lighting, the camera angle. Usually the answer is simpler: the viewer left before they even found out what the video was about.

On Reels, TikTok and Stories, the screen isn't yours — it's rented, and the rent starts on the first second. You don't buy that second with the video itself; you buy it with the first line, which the industry has shortened to a single word: the hook. What follows is which patterns actually work, which ones fall apart the moment they're borrowed word-for-word from somewhere else, and how to test your own hook without any tool at all.

Why it's the first three seconds you're losing, not the video

On a website, the visitor already made a decision — they clicked a link, they arrived on purpose. A video in a feed is a different game: nobody chose it, it just landed in front of a thumb that was already moving. If the thumb doesn't stop, the video might as well not exist, no matter how well it's made.

The platforms themselves prove the point: on skippable in-stream ads, YouTube gives the viewer a skip button after 5 seconds — as its own help documentation describes it. That's a grace period a platform deliberately grants to paid advertising. Organic Reels content doesn't get that grace period at all — there's no button, just a thumb, and it usually decides in under a second.

The conclusion is simple: the opening line isn't an introduction, it's a headline. Write it last, once the body of the video is finished, so the promise matches exactly what's inside.

Four patterns that actually work

1. The question

Weak version: a generic "Did you know..." that could open any video, which means it belongs to none of them. Nobody was actually asking that question, so it doesn't grab anyone.

Working version: ask the specific question your viewer already had this week. For a car workshop, "do you own a car?" opens nothing; "why does air conditioning wear your engine out faster in summer?" is a question that person already has — and they'll stay to hear the answer.

2. The contradiction

Weak version: "Everyone thinks..., but" without saying who — it doesn't work because a contradiction needs a real belief to push against first.

Working version: name the specific thing your customer actually believes, then flip it. If the assumption is "the more expensive cream works better," the rest of the video has to show why that's not always true — immediately, in the second line, not three minutes in.

3. The number

Weak version: "most people don't know that..." — with no quantity, the viewer can't judge how long this takes or what shape the payoff has.

Working version: use a real number from your own experience. "3 mistakes" only if you actually counted three. An invented number is written to sound credible, but viewers usually sense it, especially the second time they've seen the trick.

4. The loss

Weak version: "If you don't know this, you're losing money!" — followed by generic tips. The loss was promised, never shown, and once a viewer feels that once, they stop trusting your next video too.

Working version: use loss framing only when the loss is real and provable, and deliver on the promise in the first ten seconds, not the second minute. For example: "not signing this one document on time costs the average client around a thousand manat" — specific, checkable, explained immediately.

Borrowed hooks rarely survive the move

A line that stopped the scroll in one market often becomes an instruction the moment someone copies it word-for-word into a different one. "Wait for it" works in the original because it's short and because the phrase withholds its own payoff; copy the words without rebuilding what they withhold, and what's left is a command, not a tease.

What actually works is building the same suspense out of the audience's own way of talking — the aside that promises a payoff, the line that ends on the surprising word instead of opening with it. That isn't a translation exercise, it's closer to how people already tell each other things, which is exactly why it doesn't read like a template.

Good examples: "This happened to my neighbour — it could happen to you." "I didn't believe it either, until..." "Watch to the end, or tell me I'm wrong."

Bad examples: a flat, borrowed "This will shock you!" that reads as a template rather than a real claim; an ad-voice line with no concrete promise, like "This is the best choice!"

Before and after, by industry

• Beauty salon: "Today I'll show you..." → "Skip this step and the colour won't last half as long."
• Restaurant: "Check out our menu" → "Half the people who order this come back for it again — you'll see why in a minute."
• Real estate: "Here's the apartment" → "There's no second unit at this price in this area, and I can prove it."
• A services studio (our own example): "Meet our services" → "Most of our clients come to us without ever having fixed this one mistake."

What changes in every example isn't the delivery — the opening line is already a promise on its own, not an introduction to one.

How to test a hook yourself

You don't need a tool. Play the video with the sound off and read only the first line — on its own, out of context, does it hold you? If it doesn't, the problem isn't the edit, it's the sentence.

A second check: show someone who hasn't seen the video only the first three seconds, then ask "what do you think happens next?" If the answer comes immediately, the hook works. If they hesitate, the viewer feels the same hesitation — they just don't answer you, they leave.

Instagram's and TikTok's own analytics carry a retention graph — if the account is active, it shows exactly where the drop-off starts. If that graph falls sharply in the first three seconds, the problem isn't in the middle of the video, it's at the very start.

A hook isn't an advertisement for the video — it's the first decision itself.

If you don't have time to write a fresh opening line for every video, our copywriting and content-plan service is built for exactly that — text written separately for each platform, in the language your audience actually speaks, delivered with a monthly content plan.