The short answer
Every ad hook can be read two ways: the pattern (the form of the first line: a question, a number, a story) and the mechanism (why it holds attention: a problem bridge, a contrarian reveal, offer pressure). On 2,443 winning Meta ads, patterns are varied but 64% share one mechanism, and the strongest results come from specific pairings, like a warning opener backed by a science mechanism (28.6% in the top tier).
Two questions you can ask of any opener
When we tag the 2,443 live Meta ads on the Ad Radar board (all with $50k+ in estimated spend, snapshot of October 9, 2026), every hook gets two labels.
- Hook pattern answers what does the first line look like? A question, a number, a first-person story, a deal.
- Hook mechanism answers why does it keep people watching? Because it promises a fix for a problem, because it contradicts a belief, because the offer expires.
They are independent. "Forty bags were stolen from the carousel at Denver International in a single year." is a specific-number pattern. Its mechanism is a personal story: the number sets up the speaker's own lost bag. A different ad could open with a number and use it to set up a discount.
2,389
ads with a hook pattern and a mechanism
13
pattern values on the board
10
mechanism values on the board
64%
of winners share one mechanism
Hook patterns: the form of the opener
| Pattern | What it looks like | Ads | Reach top tier |
|---|---|---|---|
| Problem/Benefit Opener | Names the problem or the result | 714 | 9.1% |
| First-Person Story | Starts inside a moment of someone's life | 438 | 10.0% |
| Specific-Number Opener | A count, price, age or figure up front | 411 | 12.9% |
| Question Opener | Asks the viewer something | 366 | 7.1% |
| Offer-First Opener | The deal before the reason | 255 | 8.2% |
| Warning/Shock Opener | A threat, a mistake, a surprising fact | 156 | 10.3% |
| Authority Opens the Lesson | An expert or founder teaching | 44 | 11.4% |
| Before/After Reveal | Shows the change first | 4 | small sample |
| Direct Callout | Names who the ad is for | 4 | small sample |
| Bold Claim Opener | A sweeping promise | 4 | small sample |
| Statistic Opener | A data point or historical fact | 1 | small sample |
| Pattern Interrupt | Something unexpected on screen | 1 | small sample |
| Mistake Admission | The brand admits a fault | 1 | small sample |
"Top tier" is the top 10% of the board by winner score, among ads live 21 days or more (about 10% of all ads). Each pattern has its own deep page: problem/benefit, story, number, question, offer-first, warning and authority.
Hook mechanisms: why the opener holds
| Mechanism | What keeps the viewer watching | Ads | Reach top tier | Median est. spend |
|---|---|---|---|---|
| Problem-to-Solution Bridge | The promise of a fix for a named problem | 1,568 | 9.0% | $225K |
| Expert Explanation | A qualified voice is about to explain | 254 | 9.1% | $220K |
| Offer Pressure | A deal that won't last | 236 | 11.4% | $180K |
| Science/Mechanism Proof | A reason how or why it works | 158 | 9.5% | $230K |
| Personal Story | Wanting to know what happened next | 136 | 11.0% | $273K |
| Contrarian Reveal | A belief is about to be overturned | 32 | 21.9% | $258K |
| Fear of Loss | Something is at risk | 5 | small sample | |
| Identity Callout | "This is about people like me" | 4 | small sample | |
| Curiosity Gap | An open question with no hint of the answer | 4 | small sample | |
| Comparison Test | Two options about to be judged | 2 | small sample |
The contrast with the pattern table is the point. Patterns are spread across seven common types. Mechanisms are not: almost two thirds of all winners hold attention the same way, by promising a fix. That's why ads that look different in the first second can still feel identical by second ten. The rare mechanisms, led by the contrarian reveal, are where the room is. See contrarian hooks.
Where the two axes meet
The interesting data is in the pairings. Here are the combinations with at least 20 ads.
| Pattern + mechanism | Ads | Reach top tier | Median est. spend |
|---|---|---|---|
| Warning + Science/Mechanism Proof | 21 | 28.6% | $429K |
| First-Person Story + Offer Pressure | 34 | 26.5% | $331K |
| Specific Number + Expert Explanation | 39 | 20.5% | $236K |
| Specific Number + Science/Mechanism Proof | 31 | 16.1% | $296K |
| Specific Number + Problem Bridge | 293 | 11.9% | $275K |
| First-Person Story + Personal Story | 115 | 11.3% | $265K |
| Problem/Benefit + Problem Bridge | 622 | 8.8% | $209K |
| Question + Problem Bridge | 253 | 8.7% | $237K |
| First-Person Story + Problem Bridge | 212 | 6.1% | $219K |
| Warning + Problem Bridge | 92 | 6.5% | $238K |
| Offer-First + Expert Explanation | 50 | 4.0% | $238K |
| Question + Science/Mechanism Proof | 55 | 1.8% | $150K |
A caution before reading too much into the top rows: they are concentrated. Story + offer pressure is 14 of 34 ads from one luggage brand (Nobl), which also holds 6 of its 9 top-tier ads. Number + expert explanation is mostly two pet supplement brands. The direction is useful, but some of what you see is one advertiser's skill.
Three lessons hold across brands:
A warning needs a reason. Warnings paired with a science or mechanism explanation reach the top tier 28.6% of the time. Warnings that simply bridge to a product reach 6.5%. The same opening line performs very differently depending on whether the next lines explain why.
Turmeric for joint pain.
- Est. spend
- $463K
- Days live
- 399
- Format
- Video, 173s
Four words that sound like a product category, then a doubt ("Is it really as effective as people say it is, or is it all just hype?") and a credential. The mechanism, explaining why most versions fail, carries the ad. Claims in the script are the advertiser's responsibility.
A story can carry an offer. First-person stories that bridge straight to a problem and product do worst of the large groups (6.1%). Stories that carry an offer inside them (the speaker found it on sale, got it as a gift) do far better.
It's time to retrieve all those files you thought you lost.
- Est. spend
- $2.5M
- Days live
- 1442
- Format
- Video, 57s
Live almost four years. The opener frames a personal loss; the headline carries the 30% off sale. Pattern and mechanism do different jobs.
A question and a science lesson don't mix. Question openers that lead into a mechanism explanation reach the top tier 1.8% of the time (1 of 55). A question asks the viewer to think, and a mechanism asks them to learn. Doing both in the first seconds asks too much.
How to use both axes when you write hooks
- Tag your last 20 ads twice. Pattern and mechanism. If 15 share the problem-to-solution bridge, your creative isn't as diverse as it looks.
- Change the mechanism, not only the first line. New wording on the same mechanism is a variation. A new mechanism is a new concept. More on that in ad concept vs variation.
- Pair warnings with a reason and numbers with an expert. Those pairings lead the board.
- Don't stack a question on a lesson. If the body teaches, open with a number, a warning or a story.
- Read competitor ads on both axes. A competitor running ten "different" hooks on one mechanism is easier to beat than it looks.
The hooks pillar has 60 real openers organized by pattern. On the Ad Radar board both tags are filterable, and the list_taxonomy tool in the Ad Radar MCP server returns every value if you want Claude to tag your own ads the same way.
Figures marked as estimated spend come from Ad Radar's model of engagement on public Meta Ad Library ads. They are estimates, labeled as such, and are best used to rank ads against each other.