Reels Ad Hooks: How to Write the First 3 Seconds That Stop the Scroll

A Reels ad hook is the first one to three seconds of the video: the opening frame, first spoken line and on-screen text that make someone stop scrolling. Write it by naming a specific viewer, problem or result straight away, showing the product or problem in frame one, and adding a text line that works on mute. Then measure each hook with hook rate.

What is a hook in a Reels or Stories ad?

The hook is everything that happens before a viewer decides whether to keep watching. In full-screen vertical placements, that decision is fast. People flick through Reels and Stories with their thumb, and an ad that opens with a logo, a slow fade or a wide shot of nothing in particular is gone before it starts.

A hook has three layers that should work together:

• The opening frame. What is on screen at second zero. It is effectively the thumbnail, and it must make sense with no sound and no context.
• The first spoken line. What the person on camera, or the voice-over, says first. Spoken hooks work best when they start mid-thought, without a greeting.
• The on-screen text. A short line, usually near the upper middle of the frame, that carries the hook for people watching on mute.

The rest of the ad (the demo, the proof, the offer) only matters if the hook earns the next few seconds. That is why it is worth testing hooks separately from everything else, and why hook rate is one of the first numbers to check on any video ad.

Why do the first three seconds matter so much?

Meta counts a 3-second video play as a basic sign that someone stopped. Ads that hold attention tend to be shown more often and more cheaply, because Meta's ad auction favours ads people are likely to engage with. A weak opening therefore costs you twice: fewer people see the message, and each impression can cost more.

There is also a practical reason. Most of a video ad's impressions are short. If the product and the reason to care don't appear until second eight, most viewers never see them. Front-loading the point is not a style preference, it is the only way the majority of your audience gets the message at all.

What are the main types of ad hooks?

Most hooks that work fall into a small set of patterns. Each pattern below comes with example lines across different product categories. They are illustrations to adapt, not lines to copy word for word, and every claim in a real ad must be one you can support.

1. The call-out

Name the viewer so they recognise themselves. It filters for the right people and repels the wrong ones, which is a good thing.

• Skincare: "If your foundation looks patchy by noon, this is for you."
• Running shoes: "Flat feet and shin pain after every run?"
• Pet products: "Dog owners with a puller, watch this."
• Baby products: "New parents doing night feeds alone, this one's for you."

2. The visible result

Open on the outcome, then explain how you got there. Works for anything with a before-and-after or a satisfying end state.

• Cleaning product: open on a spotless hob, then cut back to the greasy version.
• Hair care: open on glossy, smooth hair in daylight, then "Three weeks ago it looked like this."
• Home organisation: open on a neat drawer, then "This took four minutes."

Keep results honest. Meta's policies restrict before-and-after images in some categories, such as weight loss and certain health claims, so check the advertising standards for your category first.

3. The bold, supportable claim

A confident statement the rest of the video proves.

• Kitchenware: "This is the only pan we've needed for two years."
• Coffee: "Café-strength coffee in 90 seconds, no machine."
• Luggage: "Everything for a week in carry-on. Here's how."

4. The myth or mistake

Correct something the viewer believes or does. Curiosity does the rest.

• Skincare: "You're applying sunscreen too early in your routine."
• Supplements: "Taking magnesium at breakfast? Here's why some people switch to evening."
• Fitness: "Stop stretching before your run. Do this instead."

5. The question

A question the viewer instantly answers in their head. Avoid yes-or-no questions where the honest answer is "no, not really".

• Jewellery: "What do you buy someone who has everything?"
• Tea: "Why does your green tea taste bitter?"
• Mattress or bedding: "Waking up hot at 3am?"

6. The pattern interrupt

Something visually unexpected in frame one: a product dropped into water, a fast zoom, an odd angle, a hand slapping the product on a table, a sudden colour change. It buys a second of attention, which you then spend on the message.

• Water bottle: thrown across the room and caught, "Still not one dent."
• Phone case: dropped face-down onto tiles in slow motion.
• Candles: the room lights switched off and the candle lit in one cut.

7. The demonstration

Start the demo immediately, with no introduction. Works especially well for products whose benefit is visible.

• Kitchen gadget: an onion diced in two seconds, mid-action at frame one.
• Activewear: a squat in leggings, held towards the camera to show they stay opaque.
• Hair tool: one pass of a straightener on a frizzy section.

8. The customer's words

Open with a real review or the way customers actually talk about the product. Customer language often beats brand language.

• Footwear: "I wore these for a 12-hour shift and forgot I had them on."
• Herbal products: "My mum asked what I'd changed. It was this."
• Bags: "I get stopped about this bag every single week."

Only use real reviews, with permission where needed, and don't edit them into claims the customer did not make.

9. The founder or maker

A face and a reason. Builds trust for small brands without a known name.

• Ayurvedic brand: "My grandmother made this oil by hand. We still do."
• Candle maker: "I pour every one of these myself, here's why it takes 48 hours."

10. The comparison

Put the product next to the thing it replaces. Compare against a category or habit, not a named competitor.

• Razor subscription: "This costs less than a coffee a month. Your current razor doesn't."
• Reusable product: "One of these replaces about a year of paper towels."
• Eyewear: "Same lenses as the optician. Here's what you pay instead."

Any number in a comparison must be true for your product and your prices.

11. The list or countdown

"Three things" promises a short, finite payoff, so people stay to the end.

• Skincare: "Three things I stopped doing that cleared my skin."
• Fashion: "Three ways to wear one linen shirt."
• Home decor: "Three cheap fixes that make a rented flat look expensive."

12. The offer up front

For warm audiences and sale periods, leading with the deal can be the hook.

• "Our festive sale is live, and this set is the one people buy twice."
• "Free shipping ends Sunday. Here's what's worth getting."

Offer-led hooks tend to work best on people who already know you. For cold audiences, lead with the problem or result and bring the offer in later.

How do you build the opening frame?

The words get most of the attention, but the opening frame does a lot of the work. A few rules that hold across categories:

1. Show the product, a face or the problem in frame one. Not a logo, not a blank background, not a slow establishing shot.
2. Start mid-action. Trim the first half-second of every clip where someone is getting ready to speak or reaching for the product.
3. Put the text hook in the safe area. In Reels and Stories, the profile name, caption and buttons sit over the top and bottom of the frame. Keep text and faces in the central area. Our Meta ad sizes guide covers the 9:16 dimensions and safe zones.
4. Make it readable at a glance. One short line, large, high contrast, on screen for long enough to read twice.
5. Design for silent viewing. Burn in captions for the spoken lines. Many people watch with the sound off, and Stories are often skipped quickly.
6. Look native. Footage that looks like it was shot on a phone by a real person often holds attention better in Reels than a polished studio clip, though this varies by brand and category. Test both.

How do you measure whether a hook is working?

Use hook rate as the main hook metric, and hold rate to see whether the rest of the video keeps people.

Hook rate = 3-second video plays ÷ impressions × 100

Example: an ad gets 40,000 impressions and 11,200 three-second plays. Its hook rate is 11,200 ÷ 40,000 × 100 = 28%.

Hold rate = ThruPlays ÷ 3-second video plays × 100

ThruPlay is Meta's count of plays to 15 seconds, or to the end if the video is shorter. Example: the same ad has 2,800 ThruPlays, so its hold rate is 2,800 ÷ 11,200 × 100 = 25%. The glossary entry on hold rate explains the variations buyers use.

Read the two together:

What is a good hook rate?

There is no universal benchmark. Hook rate varies widely by placement, audience, category, video length and how Meta counts plays in different placements. The useful comparison is against your own account: the median hook rate of your last 10 to 20 video ads in the same placement mix. Treat a hook that clearly beats that median as a winner, and one well below it as a hook to replace.

Hook rate is a means to an end, though. A sensational opening can win attention from people who will never buy. Always check that a high-hook-rate ad also produces click-through and purchases at an acceptable cost per result before you call it a winner.

How should you test hooks?

The cleanest hook test changes only the opening.

1. Pick a proven body. Use a video whose middle and end already convert, or a new concept you believe in.
2. Write three to five hooks from different patterns above, for example a call-out, a visible result and a myth.
3. Edit one version per hook, with the same body, the same music and the same end card.
4. Launch them together in one ad set on a testing budget, so they run against the same audience at the same time. The 48-hour creative testing framework shows one way to structure this.
5. Read hook rate first, after a few thousand impressions each, then wait for cost per purchase before declaring a final winner.
6. Log the winning pattern, not just the winning ad. If call-outs keep beating questions for your brand, that is worth more than any single result.

Meta's delivery system will not give every ad an equal share of impressions. That is normal. If one variant gets very little delivery, it has not lost, it has not been tested yet. Give it a fair chance in a later round or in its own ad set.

Worked example: three hooks, one body

Say a skincare brand has a 20-second video that demonstrates a moisturiser. They test three openings on the same body:

These numbers are hypothetical. The myth hook wins on hook rate but loses on hold rate and cost per purchase: it stopped people who weren't interested in the product. The call-out is the one to scale. The next test might pair the call-out line with the absorbing-cream visual as the opening frame.

What are the most common hook mistakes?

• Opening on the logo. Brand recognition matters, but put the logo at the end or in a corner, not in frame one.
• Saying hello. "Hi everyone, today I want to show you..." spends the hook on nothing.
• Burying the product. If the product doesn't appear until second six, many viewers never connect the ad to what you sell.
• Text in the wrong place. Lines hidden under the caption or the buttons are invisible in Reels and Stories.
• Clickbait that the ad doesn't pay off. It raises hook rate and lowers everything else, and it can earn negative feedback.
• Claims you cannot support. Hooks are where exaggeration creeps in. Meta reviews ads against its advertising standards, and unsupported health, weight-loss or income claims get ads rejected.
• Changing everything at once. If the hook, music, body and offer all differ between two ads, you won't know which change made the difference.

How do hooks fit into a wider creative plan?

Hooks are the fastest lever to pull, but they sit on top of the angle, which is the reason someone buys. A good process keeps the two separate. Pick the angle from your library of ad angles, then write several hooks for each angle and test them. When an angle starts to tire, a new hook on the same angle is often the cheapest refresh, as the guide to ad fatigue explains.

For written copy, the same thinking applies to the first line of the primary text. The Facebook ad copy formulas guide covers openers for feed ads, and the ad copy checker can score a line before you spend on it.

If you want hook variants without editing every version by hand, Fly ROAS drafts several openings per angle from your real product photos and brand voice, and its 48-hour tests show which one earns the spend.

Reels hook checklist

• Product, face or problem visible in the first frame
• First spoken line starts mid-thought, with no greeting
• One short on-screen line in the safe area, readable on mute
• Captions burned in for every spoken line
• Three or more hook variants per body
• Hook rate and hold rate columns set up in Ads Manager
• Winners judged on cost per purchase, not attention alone

Open with the point

The best hooks don't trick people into watching. They tell the right people, in one second, that this video is about them. Name the viewer or the problem, show the product early, write for silent viewing, and let hook rate tell you which openings to keep.