There is no typical number of Meta ads, and no typical lifespan. On October 26, 2026, ten US skincare brands that sell direct were running between 10 and 830 live ads each, with a median of 119, and their live ads had been running a median of 2.7 to 25.5 days. Budget sets how many ads a brand runs; testing sets how long each one lives.
I'm Caner Moral, founder of AdRiseLab. I have managed Meta ads for five years, across beauty, mobile apps and SaaS, ecommerce, lead generation and local businesses. Over that time I have looked at the Meta ads of 150 brands in those five industries, and this post puts what I saw next to one day of public data: the live US ads of ten skincare brands, pulled from the Meta Ad Library. The brands are not named here; the method and its limits are at the end.
How many Meta ads brands run at once
The ten brands had 1,915 live ads between them. The smallest ran 10, the largest 830, and the median was 119. Six of the ten sat between 73 and 172; the other two big accounts ran 306 and 830.
Across the 150 brands I have looked at, the spread is the same in every industry. I have seen brands run 10 ads and brands run 250, and at large companies I have seen more than 1,000 running at once. Nothing in this post surprises people more than that number. There is no meaningful average, because the count follows the budget: every ad needs enough spend to produce a result you can read.
So copying a competitor's ad count tells you very little. A brand spending thousands a day can feed hundreds of ads; one spending $50 a day cannot feed more than a handful and still learn anything. To work out how many ads your own budget can test at once, use the creative testing budget calculator, and for a planning view by budget see how many Meta ad creatives you need.
How long live ads had been running
The Ad Library shows when each live ad started, so you can see how long it has been running so far. Across the ten brands, the median live ad had been running between 2.7 and 25.5 days. That is the age of the ads that were still live on the day we looked, not the full lifespan of an ad: a live ad keeps ageing until someone switches it off.
The list is cut at each brand's newest 50 live ads, so for the seven brands with more than 100 live ads the figure is a minimum. Two brands showed their full list. Their live ads had a median age of 7.0 and 17.8 days, and none of them was older than 38 days.
That range matches what I see in my own accounts. In a test campaign, most ads live somewhere between about 3 and 25 days. The ones that win do not stay there. I move them into a separate campaign, where a winner can keep running for two to three months.
There is no fixed day on which an ad stops working, which is why the decision belongs to your results rather than the calendar. The signs to watch are in our guide to spotting creative decay and the creative fatigue explainer.
Why there is no weekly number of new ads
On the day we looked, the newest ads were very new. At four brands, all 50 visible ads had started within the past seven days, and the two full lists showed 1 and 30 new ads that week.
They also arrive in batches. Nine of the ten brands created 12 or more ads within a single ten-minute window. At the largest account, the newest 50 ads were all created within five minutes on October 23.
People often ask me how many new ads a brand should launch each week. There is no such number. Brands launch when a test finishes and the data tells them what to try next, so launches come in bursts. Testing is what decides the rhythm, and the creative testing framework covers how to set one up.
One visual, many messages
An ad count is not a creative count. Among the ads each brand started that week, a single headline ran in 1 to 5.6 ads on average, depending on the brand. One brand created 42 ads in nine minutes on October 25: seven offer headlines, six ads each.
I see the same pattern in the accounts I manage. Some brands run one visual with different messages and hook lines across different campaigns. The visual stays the same; the message is what gets tested.
When I look at the creatives that stay live longest, the message matters most, and the hook matters as much as the message. A strong visual with a weak first line rarely lasts. Our collection of hook formulas for the first three seconds and the hook rate explainer show how to write and measure that first line.
Video or image
Video share of live ads ranged from 21% to 78% across the ten brands, with a median of 56%. Four of the ten ran more images than videos.
This shows what brands run. How each format performs is a separate question, covered in our analysis of static versus video Meta ads.
What changes before Black Friday
Around Black Friday, ad counts go up on average and the messages turn to discounts. Our late-October pull already showed it starting: one brand had created 22 holiday-themed ads in four minutes on October 23.
Black Friday falls on November 27, 2026. If you are still building your plan, the 28-day BFCM calendar sets out when to lock audiences and creative.
How to use this in your own account
These numbers describe how brands work, and they make poor targets. The structure I use across all five industries has four steps:
- 1.Research rivals in the Ad Library: which messages they repeat, and how long each ad has been running.
- 2.Write several messages and hooks for one strong visual, and test them together in one test campaign.
- 3.Keep or pause each ad on your own results, not on a fixed number of ads or days.
- 4.Move winners into a separate campaign, and keep the test campaign supplied with new messages.
A long-running ad in someone else's account is a hint, not proof. The Ad Library shows no spend or results for commercial ads outside the EU, so an ad can run for months on a small budget. Our guide to what the Meta Ad Library shows and what it can't covers those limits.
Where AdRiseLab fits
AdRiseLab covers each of the four steps in one workspace. Its competitor research searches the Meta Ad Library by brand or keyword, tracks rival brands and shows how long each ad has run. Creative generation turns a product URL into image ads and copy, with every batch built from different hook types and angles, and also makes video and creator-style ads, so you can test several messages on one visual without a design team.
When you confirm a launch, AdRiseLab creates the campaign, ad set and ads in your Meta account, and its reports show which message won. Plans start at $39 a month; see pricing.
Method and limits
Data: Meta Ad Library, United States, active ads only. Ten US skincare and beauty brands that sell from their own websites, chosen because they advertise actively, not at random. Pulled on October 26, 2026 at 14:46 UTC. Brands are shown as letters A to J, ordered by live ad count.
For each brand we could see the number of live ads and the start date of its newest 50 live ads: 460 of the 1,915. Video and image counts come from the Ad Library's media type filter, where counts above 100 are rounded by Meta and carousels are counted by their media.
| Brand | Live ads | Median age of live ads | Video share | Largest 10-minute batch |
|---|---|---|---|---|
| A | 10 | 17.8 days | 60% | 4 |
| B | 50 | 7.0 days | 78% | 12 |
| C | 73 | about 10.3 days | 21% | 22 |
| D | 112 | at least 3.8 days | 52% | 29 |
| E | 116 | at least 17.6 days | 33% | 12 |
| F | 122 | at least 10.8 days | 27% | 42 |
| G | 124 | at least 7.0 days | 28% | 16 |
| H | 172 | at least 25.5 days | 59% | 19 |
| I | 306 | at least 3.7 days | 60% | 16 |
| J | 830 | at least 2.7 days | 62% | 50 or more |
What this data cannot tell you: spend, impressions and results are not shown for commercial ads outside the EU, so nothing here says which ads made money. Age is time live so far, not total lifespan. Ten brands in one category do not represent every advertiser, and late October is the start of the holiday season.
The practitioner observations come from the Meta ads of 150 brands I have looked at over five years, across beauty, mobile apps and SaaS, ecommerce, lead generation and local businesses. They are experience, not a measurement: every number in the charts and the table comes from the ten-brand pull.
Sources
Meta Ad Library, United States, active ads, accessed October 26, 2026 at 14:46 UTC, with the media type filter for video and image counts. Practitioner observations: Caner Moral, AdRiseLab, from 150 brands reviewed over five years.
Related reading
What the Meta Ad Library shows and the 9 things it can't, how many Meta ad creatives you need, the creative testing framework and the creative testing budget calculator.