The short answer
Specific-number openers have the highest top-tier rate of any common hook pattern on our board: 12.9%, against about 10% across 2,443 winning Meta ads, and the highest median estimated spend ($274K). The kind of number matters more than its precision: list counts ("Five rules...") reach 17.8%, while measurements and time spans sit below average.
The numbers on number hooks
Of the 2,443 live Meta ads on the Ad Radar board on October 9, 2026 (all with $50k+ in estimated spend), 411 open with a specific number, from 200 brands. They beat every other common hook pattern on both of the measures we track.
| Hook pattern | Ads | Reach top tier | Median est. spend |
|---|---|---|---|
| Specific-Number Opener | 411 | 12.9% | $274K |
| Warning/Shock Opener | 156 | 10.3% | $250K |
| First-Person Story | 438 | 10.0% | $231K |
| Problem/Benefit Opener | 714 | 9.1% | $209K |
| Offer-First Opener | 255 | 8.2% | $183K |
| Question Opener | 366 | 7.1% | $219K |
"Top tier" is the top 10% of the board by winner score, among ads live 21 days or more. About 10% of all ads reach it. Specific-number openers also have more ads above $1M in estimated spend than average (11.9% vs 10.8%) and run extra variants more often (32% have two or more, against 29% for the board).
Which numbers do the work
Not every number is equal. We sorted every opener that contains a number (705 ads, video first sentences and image headlines) by what the number describes.
| What the number is | Ads | Reach top tier | Median est. spend |
|---|---|---|---|
| A big count ("50,000 families", "thousands") | 28 | 21.4% | $231K |
| A list count ("Three signs", "Five rules") | 45 | 17.8% | $337K |
| A price or discount ("$2,300", "50% off") | 128 | 12.5% | $229K |
| A year ("since 1914", "of 2024") | 20 | 10.0% | $248K |
| An age ("over 50", "22-year-old") | 51 | 9.8% | $159K |
| A time span ("seven days", "90 days") | 159 | 8.8% | $209K |
| A measurement ("50 grams", "5'10") | 30 | 6.7% | $182K |
Two groups clearly lead. List counts combine a high top-tier rate with the highest median estimated spend of any group ($337K). Big counts reach the top tier at more than twice the board rate, though 28 ads is a small sample.
Time spans are the most common kind of number and sit below average. "For four weeks" and "in 90 days" have become so standard in supplement and pet ads that the number no longer surprises anyone.
Precise isn't the same as specific
There is a common belief that odd, precise numbers (47%, 263 pounds) beat round ones because they sound measured. Our board doesn't support it.
8.4%
top-tier rate for precise figures (155 ads)
10.5%
top-tier rate for round or small figures (314 ads)
What wins is a number attached to something concrete the viewer can picture or check. "Forty bags" is round. It still works because it is forty bags, from one carousel, at one airport, in one year.
Forty bags were stolen from the carousel at Denver International in a single year.
- Est. spend
- $4.1M
- Days live
- 126
- Format
- Video, 85s
- Variants
- 6
A round number made specific by place and time. The script then turns into a first-person story ("Mine never showed up"), so the statistic sets up the stakes and the story carries them.
The list count: a promise with a length
A list count tells the viewer exactly what they will get and how long it will take. It also creates a reason to stay to the end, because nobody wants to stop at rule three of five.
Five brow rules every woman over 50 needs to follow.
- Est. spend
- $816K
- Days live
- 131
- Format
- Video, 82s
- Variants
- 3
Count, topic and audience in nine words. Rule one ("stop using brow products made for 20-year-olds") sets up the product as the answer to the list.
Other list openers on the board follow the same structure: number, noun, audience. Our hold rate analysis found the countdown is also one of the structures that keeps viewers watching after the hook.
The time span: make it a test, not a promise
Time spans underperform when they promise a result ("results in 30 days"). They do better when they frame the ad as a test the speaker ran.
I tried a meal delivery service for the past seven days, and here's what I think.
- Est. spend
- $3.5M
- Days live
- 145
- Format
- Video, 51s
- Variants
- 2
Seven days is the length of the experiment, not a promised outcome. The viewer hears a review, not a claim.
Prices and discounts
Price numbers are the second most common group (128 ads) and land above average at 12.5%. Most are discount headlines on image ads aimed at people who already know the brand. Hair Envy has run "Hair Envy 65% OFF TODAY!" for 1,530 days with 12 variants (est. $1.37M). That is not a clever hook. It is a number the returning buyer was waiting for, which is a different job from stopping a stranger. When to open on the deal is covered in offer-first hooks.
Ages and counts of people
Ages work best when they belong to a character, not to the audience. "Lulu is a 22-year-old mutt." (The Farmer's Dog, est. $387K, 104 days, top tier) puts a remarkable age on a dog the viewer can see, in a 10-second ad. "Women over 50" in an opener is a callout, which is useful targeting, but the number itself isn't the hook.
Counts of people do the job of social proof in the first line.
Trusted By 250,000+ Athletes
- Est. spend
- $332K
- Days live
- 306
- Format
- Image
- Variants
- 3
An image headline that is nothing but a count of customers. It has run for 306 days, which suggests the number keeps doing its job long after launch.
Cozy Hoome stacks two numbers about the product itself: "This is a smart flash drive for your phone that holds 250,000 photos and 500 hours of HD videos." (est. $1.21M, 473 days). The numbers turn an abstract spec (storage) into things the viewer owns.
How to write a number hook
- Find the number you already have. Customers served, units sold, years in business, steps in your method, price of the alternative. Every brand has one.
- Prefer a count or a list over a time span. List counts reach the top tier at 17.8% on our board. Time spans sit at 8.8%.
- Attach the number to a noun the viewer can picture. Bags, shoes, grams, dollars. "Five rules" is better than "five ways to improve".
- Don't fake precision. 263 doesn't beat 250 on our board. Use the real figure, rounded the way a person would say it.
- Use the price of the alternative. "$2,300 for that." (Pop On) and "$5,000 on a bed" (Essential Sleep Hacks) make the product's price look small without stating it.
- Be able to prove it. Any statistic in an opener is a claim. If you can't source it, don't run it.
Number openers pair well with the structures in video ad hook formulas, and they are a good fix for weak question hooks. For more numbered openers by niche, the Ad Radar board lets you filter by hook pattern and sort by estimated spend.
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.