Ad Hooks: 10 Openers That Grab Attention
10 min read · AstraLoop Studio
You can have the best product on the market, the sharpest offer and flawless editing. If the first three seconds don't stop the thumb, none of it gets seen. The hook is that opening beat, visual or verbal, that decides whether someone stops or keeps scrolling. It's not a cosmetic detail: it's the single variable that moves creative performance the most.
The numbers back this up. According to Meta's internal data, 65% of people who watch the first 3 seconds of a video make it to at least 10 seconds. The best-performing ads post a thumb-stop rate (the share of people who stop) of 30-40%, against 15-20% for the average. In other words: improving your hook by 10% can double your effective reach without spending a single extra euro on budget.
In this article you'll find 10 hook formulas, split between visual and verbal, with concrete examples you can adapt to your product today. At the end we cover how to fit them into your video structure and how to test them without burning through budget.
Why the hook matters more than everything else (especially in 2026)
Until recently, a good concept was enough to keep a campaign running for weeks. Not anymore. With Meta's delivery system nicknamed Andromeda, the algorithm reads the structure of the hook and evaluates the first three seconds of a video separately to predict which audience to show it to. In practice, the hook doesn't just stop the user - it's a signal the AI uses for targeting.
There's more. A single concept today "burns through" its audience in 2-3 weeks, versus 6+ in the pre-Andromeda era. That means one thing: you need many different hooks, not one brilliant opener. If you want to understand in depth how the landscape has shifted, we dedicated a full piece to what changes with Meta Andromeda for creative. Here we're focusing on the raw material: the openings themselves.
One last technical note. 85% of videos on Facebook are watched without sound. So every verbal hook has to work on a muted screen, with large, high-contrast on-screen text. That's not optional, it's the baseline rule. If you haven't already, read why subtitles make a difference in video ads.
5 visual hooks that stop the thumb
A visual hook works before the brain even reads a word. It's the first frame, the movement, the color, the face. Here are the five most reliable formulas.
1. The pattern interrupt (break the pattern)
The feed is a uniform stream of predictable content. An out-of-place element creates a micro-friction that stops the scroll: an unusual angle, a quick zoom, an object that has nothing to do with the scene. The brain notices the anomaly and pauses to make sense of it.
Adaptable example: for a face cream, open with the cream being spread on a mirror instead of a face, with the hand then "wiping" a circle to reveal the reflection. Just strange enough to stop the scroll, while staying on-brand for the product.
2. The result first (reversed before/after)
Show the final result, the transformation, or the "after" in the first three seconds, then rewind to explain how you got there. It works because the benefit is immediate and viewers want to know how to replicate it.
Adaptable example: for a kitchen tool, first frame on the perfect finished dish, then a hard cut back to the "before" with the raw ingredient. Curiosity about the process keeps the viewer glued to the screen.
3. The close-up face talking to you
A human face, looking straight into the camera in tight close-up, triggers more attention than any product shot. If the person seems to be talking directly to you (UGC format, natural light, a home setting), the effect multiplies. It's the engine behind why UGC ads work so well.
Adaptable example: first frame on someone with wide eyes and a hand over their mouth, big caption reading "I didn't believe it until I tried it." Expression + text = two hooks in one.
4. The hypnotic gesture or texture (visual ASMR)
Some movements are simply hypnotic: fabric flowing, a liquid pouring slowly, a hand slicing something cleanly, a label being peeled off. The brain follows the motion by default. These are silent hooks, but extremely powerful ones for physical products.
Adaptable example: for a piece of clothing, open with a zoom on the fabric's texture as it's stroked or folded, then pull back to reveal the garment being worn. Make people want the "touch" before they even see the full picture.
5. Giant text as the first frame
Sometimes the strongest visual is a single word. An opening frame with just one sentence, huge, high-contrast, that promises or provokes. It works on a muted screen by definition, and it's extremely fast to produce and test.
Adaptable example: solid background, white text on black: "I stopped buying X. Here's why." Then the demo kicks in. It's the cheapest hook to iterate on: you only change the sentence and you have a new version to test.
5 verbal hooks that hook people in the first words
The verbal hook is the first line, spoken or written. It has to create a tension that only resolves if you keep watching. One caveat: verbal doesn't mean "audio only." The line should always be captioned on screen too. Here are the five most solid levers.
6. The question that names the problem
A well-calibrated question makes people who recognize themselves in it mentally raise their hand. The secret is specificity: the more precise the problem, the more the person living it feels called out.
Adaptable example: "Do you wake up already tired even after 8 hours of sleep?" for a supplement. "Does your sofa always have that stain that won't go away?" for a cleaning product. This lever works best when you know your audience well: it helps to think through the five levels of customer awareness to figure out which problem to name.
7. The contrarian statement (against common sense)
Opening by challenging a widespread belief creates productive friction. Viewers think "wait, how's that possible?" and stay to find out. It has to be a defensible position, not an empty provocation.
Adaptable example: "Washing your hair every day is mistake number one." Or: "Your morning coffee isn't waking you up, it's slowing you down." Then the explanation and the product follow. The line itself creates the tension - nothing else is needed.
8. The surprising number or statistic
A precise, unexpected figure is an extremely powerful hook because it's concrete and credible. Round, specific numbers beat vague statements. "Many people" stops no one; "9 out of 10 people" does.
Adaptable example: "I saved €340 in three months by doing just one thing." Or: "78% of people who try it buy it again within a month." The number has to be real: inflating it backfires the moment you try to scale.
9. The confession or the behind-the-scenes secret ("no one tells you this")
People are drawn to insider information, to what happens behind the curtain, to the thing "the industry doesn't want you to know." It opens a curiosity loop that only closes if you keep watching.
Adaptable example: "I've worked in this industry for 10 years and we used to keep this product to ourselves." "No store will ever tell you this about jeans." It's a classic persuasion lever, and it pairs well with copywriting frameworks like AIDA and PAS to structure what comes after the hook.
10. The mistake to avoid (fear of getting it wrong)
Naming a common mistake triggers the fear of already being the one making it. It's irresistible: no one wants to find out they've been doing something wrong for years without realizing it.
Adaptable example: "You're using this product the wrong way (everyone does)." "3 mistakes that ruin your shoes without you noticing." It works great as a list format because it also promises the structure of the content.
Want to go from a handful of creatives to a steady stream of tested hooks without overloading your team? Tell us about your product and we'll show you how to set up production and testing with AI.
How to fit the hook into your video structure
A brilliant hook isn't enough if what follows doesn't hold up. The hook opens a loop, and the rest of the video has to close it in an orderly way, or the viewer leaves anyway. The sequence that works almost every time is: hook (0-3s) → context/problem (3-8s) → solution and demo (8-20s) → proof and CTA (20-30s).
The critical point is consistency between the hook and the body. If you open with "3 mistakes with your shoes," the three mistakes need to show up right away, not after thirty seconds of preamble. The opening's promise is a contract: keeping it within a few seconds is what keeps retention high. We have a deeper dive on the structure of video ads that convert and a more specific one on reels that stop the scroll, where the pacing is even tighter.
A practical rule for on-screen text: a maximum of 7 words per frame, large font, high contrast, positioned in the safe zone (away from edges and platform UI). The hook has to be readable in half a second, not "read carefully."
How to test hooks without burning through budget
Here's the most common mistake: changing the entire creative on every test. That way you learn nothing, because you don't know which variable moved the result. The right method is to isolate the hook.
Take a creative that's working, keep the body and CTA identical, and change only the opening. Produce 5-10 hook variants on the same concept and compare the thumb-stop rate of each. After a few cycles, patterns emerge: you'll find that for your audience, the problem-question always beats the contrarian statement, or the other way around. Those patterns become your most valuable asset.
| Metric | What it tells you | Benchmark |
|---|---|---|
| Thumb-stop rate (hook rate) | How many people stop in the first 3s | Good ≥ 30%, average 15-20% |
| Hold rate (25%/50%) | Whether the body delivers on the hook's promise | Higher is better |
| CTR | Whether the promise drives action | Depends on the industry |
In 2026, volume matters more than ever: the data shows that brands testing 20+ new ads per month achieve higher ROAS than those testing fewer than 10. You don't need to reinvent everything each time: many different hooks on the same solid concept are enough. If you want a structured method, we wrote a guide on the creative testing method for ads. And when you run out of ideas, the piece on how to find ideas for ad creatives will unblock you.
Where to find your next hook idea
The best hooks rarely come from sitting at a desk. They come from customers' real words: reviews, recurring objections, the questions people ask you in person. When a customer tells you "I thought it would be too complicated, but...", that's already the first line of a hook. Listen to the language they use and feed it back into your opening.
Producing a lot of variants is, today, the bottleneck for almost every company, and this is where AI changes the game: generating 10 hook versions, with different angles and captions, starting from a validated concept, has become fast and accessible. It's exactly the ground we work on. If you want to see how the hook fits into the bigger picture, our pillar piece on ad creative: the complete guide pulls all the pieces together, from idea to testing.
One last piece of advice, the most important one: don't fall in love with your favorite hook. The feed decides, not your taste. Put many openers into rotation, watch the numbers, and keep the ones that stop the thumb. That's the only judge that counts.
Frequently asked questions
How long should a hook be in an ad?
The hook has to grab attention within the first 3 seconds, ideally within 1.5. Meta evaluates the first three seconds of a video separately, so that's where everything is decided. If you haven't stopped the scroll in that window, the rest of the creative will almost never get seen.
What's the difference between a visual hook and a verbal hook?
A visual hook works on the first frame, the movement and the color, before the brain even reads anything: a face, a texture, a pattern interrupt. A verbal hook is the first line, spoken or written, that creates tension or curiosity. The best ones combine both: an expressive face with a strong line of text on top.
Do I need to caption the hook on screen too?
Yes, always. 85% of videos on Facebook are watched without sound, so every verbal hook has to work on a muted screen. Use large, high-contrast text, a maximum of 7 words per frame, inside the platform's safe zone.
How many hook variants should I test?
5 to 10 hook variants on the same concept, keeping the body and CTA identical to isolate the variable. In 2026, volume pays off: brands testing over 20 new ads per month post higher ROAS than those testing fewer than 10. Many different hooks on the same solid concept beat chasing the perfect opener.
How do I know if a hook is working?
Look at the thumb-stop rate (or hook rate), the share of people who stop in the first 3 seconds. Above 30% is excellent, the average sits between 15 and 20%. Then check the hold rate at 25% and 50% to see whether the body of the video delivers on the promise made in the opening.
Can I use AI to generate hooks?
Yes, and it's one of its most effective uses. Starting from a validated concept, AI can quickly produce dozens of opening variants with different angles, lines and captions, removing the production bottleneck. The key is to start from customers' real words (reviews, objections) and then let the numbers decide which hooks to keep.
If the bottleneck is producing enough variants, let's talk: we'll look at your case together and define a system to generate and test hooks at a fast pace.