The short answer
Longer than most media buyers expect. Across 2,443 live Meta ads with $50k or more in estimated spend, the median ad has been running for 120 days, and 1 in 7 has passed a full year. Winners don't get replaced every two weeks. They get new variants around them.
The short version in numbers
We looked at every ad on the Ad Radar board on October 9, 2026: 2,443 active Meta ads from 749 brands, each with at least $50k in estimated spend. For each one we took the start date from the Meta Ad Library and counted the days to today.
120 days
median time live
42%
live between 90 and 180 days
14%
live for more than a year
30 days
shortest run on the board
Two things to keep in mind. First, these ads are still running, so every figure is a floor: the final run will be longer. Second, this is a board of winners. It tells you how long ads run once they work, not how long a typical test lasts.
The full distribution
| Days live | Ads | Share |
|---|---|---|
| 30 to 59 | 172 | 7% |
| 60 to 89 | 529 | 22% |
| 90 to 179 | 1,037 | 42% |
| 180 to 364 | 367 | 15% |
| 365 to 729 | 247 | 10% |
| 730 or more | 91 | 4% |
The bulk sits between three and six months. That is roughly one season of a product's life: long enough to scale, short enough that most brands are already testing the next version.
The tail is the interesting part. 338 ads have been live for more than a year, and 91 for more than two. These are the ads a brand builds its account around.
Longer runs, more spend
Time and spend move together, which is what you would expect if brands keep feeding budget to what works. The median estimated spend climbs with each band:
| Days live | Median estimated spend |
|---|---|
| 30 to 89 | $183K |
| 90 to 179 | $203K |
| 180 to 364 | $271K |
| 365 or more | $345K |
This is why days live is one of the best public signals you have when you research competitors. A brand can launch an ad on a whim. It doesn't keep paying for it for a year on a whim.
Airport thieves are now stealing your stuff with just a pen.
- Est. spend
- $5.2M
- Days live
- 882
- Format
- Video, 45s
Almost two and a half years on the same opening line. The hook names a specific threat, then the product is the fix.
Does it change by niche?
Less than you might think. Most large niches on the board land between 110 and 155 days.
| Niche | Ads | Median days live |
|---|---|---|
| Foot | 135 | 155 |
| Digestion | 239 | 154 |
| Pet | 313 | 136 |
| Joints | 208 | 133 |
| Circulation | 138 | 119 |
| Energy | 132 | 117 |
| Skin | 224 | 114 |
| Stress and mood | 139 | 114 |
| Cleaning | 151 | 113 |
| Food and nutrition | 164 | 112 |
Foot care and digestion run the longest. Both are categories where the problem is chronic and the audience is wide, so one strong problem-aware ad can keep finding new people for months. See what gut health winners run and what foot care brands run for the ads themselves.
Does it change by format?
Here the spread is wider. Product demos last the longest on the board, AI-generated creative the shortest.
| Production format | Ads | Median days live |
|---|---|---|
| Product demo | 97 | 173 |
| Footage + voiceover | 177 | 142 |
| UGC selfie | 78 | 140 |
| Founder or expert | 145 | 128 |
| AI-generated | 77 | 97 |
A demo shows the product working, and that doesn't go stale the way a trend or a joke does. AI-generated ads reach the board, but they tend to be swapped out sooner. We look at why in AI UGC ads: what the spend data says.
Today I'm going to show you the basic techniques you need to look your very best.
- Est. spend
- $3.2M
- Days live
- 1288
- Format
- Video, 51s
Over three and a half years live. A tutorial is a demo in disguise: it shows the product doing its job.
Winners get variants, not replacements
Most winning ads on the board run as a single version (1,743 of 2,443). The rest carry two or more variants, and 28 have ten or more. When a brand finds an ad that works, the common move is to keep the body and test new openers or new cuts around it, not to kill it and start over.
That matters for how you read "creative fatigue". If your best ad drops after six weeks, the data says the answer is usually a new hook on a proven body, not a new concept from scratch. More on that in creative fatigue: how to spot it early and how to scale a winning ad.
What to do with this
- Test fast, keep winners long. A test needs days. A winner can carry an account for months. Don't apply your test cadence to your winners.
- Read days live when you research competitors. An ad live for 90 days or more with a spend estimate in the six figures is not a test. Start your swipe file there.
- Plan refreshes as variants. Budget creative time for new hooks on ads that already work, not only for new concepts.
- Be wary of quick swaps. If you replace your best ad every two weeks, you may be throwing away the asset the algorithm was still finding buyers for.
If you want to see these ads, filter by days live and estimated spend, and read the scripts, that is what the Ad Radar board is for.
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.