Every ad we build lives or dies in its first three seconds. Not the offer, not the price, not even the product — the hook. If someone's thumb doesn't pause in that window, nothing else in the ad matters, because nobody sees it.
This isn't a creative-writing exercise. It's a structural one. The strongest hooks we've tested share a specific anatomy, and most underperforming ads fail for one of a handful of predictable, fixable reasons.
01The two-layer hook structure
Every effective hook has two layers, and they happen almost simultaneously — but they do different jobs.
| Layer | Timing | Job |
|---|---|---|
| Visual hook | 0–1 second | Interrupt the scroll — movement, unexpected color, a face, a surprising action |
| Verbal/text hook | 1–3 seconds | Give a reason to keep watching — the actual "why should I care" moment |
Most ads that fail in testing get the visual layer right but skip the verbal layer, or the reverse — strong opening line over a static, unremarkable shot. You need both working together. The visual earns the pause; the words earn the next five seconds.
02Eight hook formulas that consistently test well
These aren't the only hooks that work, but they're the ones we return to across categories and platforms because the underlying psychological trigger is durable — it doesn't wear out as fast as a clever one-off line does.
03What to cut from a weak hook
When a hook underperforms in testing, the fix is almost never "make it more creative." It's usually one of these specific, boring problems:
- Logo or branding shown before the hook lands — the moment something looks like an ad, the brain's ad-blocker (real or mental) kicks in and the scroll continues.
- The hook explains the product instead of creating curiosity — "Introducing our new skincare serum" is information, not a hook. It answers a question nobody asked yet.
- Text hook takes longer than 3 seconds to read — if the on-screen text requires pausing the video to finish reading it, it's too long for a sound-off scrolling environment.
- No visual movement in the first frame — a static opening shot, even a beautiful one, doesn't interrupt a thumb mid-scroll the way motion does.
04How to actually test this
Don't test one "great" hook against a generic control. Test five genuinely different angles against the same body content — a direct callout, a shock stat, a myth bust, and two others from the list above. The angle that wins isn't necessarily the most clever one; it's the one that resonates with the specific audience segment the algorithm ends up showing it to.
Track hook rate specifically — the percentage of people who watch past the 3-second mark. A hook rate above 30% is generally a strong signal on Meta; below 15-20% usually means the hook needs to be rebuilt, not just the body of the ad.
The best-performing hook in your account six months from now probably doesn't exist yet. Hooks fatigue the same way any creative does — which is exactly why this is a formula to keep testing against, not a single line to write once and reuse forever.