Video ad hooks: 12 opening patterns
A video ad hook is the first one to three seconds, the part that decides whether anyone watches the rest. Strong hooks state a problem, a result or a surprising fact immediately, show motion or a face, and put text on screen for muted viewers. The fastest way to improve an ad is to test several hooks on the same body.
Updated
Why the hook carries most of the weight
In a scrolling feed, the default outcome for any ad is to be skipped. The hook is the only part every impression actually sees, so it has an outsized effect on everything downstream: views, clicks and, through auction quality signals, what you pay for delivery.
It is also the cheapest part to change. Keeping the body and swapping the first two seconds produces a new ad without new production, which is why hook testing is usually the first creative test worth running.
Twelve hook patterns
| Pattern | Example line | Works best for |
|---|---|---|
| Problem callout | 'If your plants keep dying, it's probably this.' | Products that fix a familiar annoyance |
| Result first | 'This took our checkout from 4 steps to 1.' | Products with a visible outcome |
| Surprising fact | 'Most ads are cropped before anyone sees them.' | Educational angles |
| Direct question | 'Still paying for three tools that do one job?' | Consolidation or switching offers |
| Demonstration | Product doing its job in the first frame, no words | Visual products |
| Comparison | 'Same coffee, half the price. Here's how.' | Price or value positioning |
| Mistake | 'Stop doing this when you write cover letters.' | Expertise-led products |
| List promise | 'Three things I'd never buy again, and one I would.' | Creator-style reviews |
| Reaction | A genuine face reacting to the product | Taste, beauty, unboxing |
| Specific audience | 'For freelancers who invoice in two currencies.' | Niche products |
| Before and after | The messy state, then the fixed state | Cleaning, organising, design |
| Story start | 'Last year I nearly shut the business down.' | Founder stories |
Rules that apply to every pattern
- Start mid-action. No logo cards, no 'Hi, I'm…' before the point.
- Put the hook in on-screen text as well as audio; many viewers start muted.
- Keep the text inside the placement's safe zone.
- Make the hook true. A hook the body does not pay off raises hook rate and lowers hold rate, and misleading hooks can breach ad policies.
- Match the hook to the audience temperature: cold audiences need the problem, warm audiences can start with the product.
How to test hooks
- Pick one ad body that is already performing reasonably well.
- Write three to five hooks from different patterns above.
- Cut each hook onto the identical body, so the opening is the only difference.
- Run them at the same time with comparable budget.
- Read hook rate first, then hold rate and cost per result. Keep the hook that wins on the business metric, not just on views.
Questions
- How long should a video ad hook be?
- One to three seconds. Platforms differ, but in scrolling feeds most people decide whether to keep watching within that window.
- What is a good hook rate?
- There is no reliable universal benchmark. Compare a new hook against your own ads in the same placement and audience.