How we measured it
The Cloud Ten ad library tracks 6,146 ads from 40 DTC brands in the Meta Ad Library, records the day each one first appeared and re-checks every day whether it is still live. 3,823 of those ads (62%) are video.
Spend and results are private, so we use days in market as the performance signal. An advertiser re-approves a live ad with their own budget every morning. Tests get switched off within a couple of weeks; an ad still running after three months has paid for itself, and an ad running after a year is a category truth.
For each brand we took its longest-running static ads and labelled the persuasion mechanism doing the main work: the one thing the ad relies on to make the sale. 276 ads carry a clear label. Each one has been live for at least 45 days, the longest for 939. 92 of them have passed 100 days.
The mechanisms behind long-running ads
| Mechanism | Ads | Share | Median days live | Longest | Brands using it |
|---|---|---|---|---|---|
| Feature callout | 66 | 24% | 74 | 939 | 20 |
| Big claim | 39 | 14% | 85 | 549 | 18 |
| Product hero | 35 | 13% | 82 | 450 | 18 |
| Offer | 31 | 11% | 87 | 149 | 10 |
| Comparison | 27 | 10% | 96 | 121 | 10 |
| UGC | 24 | 9% | 72 | 457 | 14 |
| Before and after | 22 | 8% | 77 | 457 | 8 |
| Social proof | 14 | 5% | 80 | 918 | 9 |
| Lifestyle | 9 | 3% | 78 | 457 | 6 |
| Chat screenshot | 6 | 2% | 80 | 219 | 4 |
Three findings stand out.
1. Plain beats clever
Feature callouts, big claims and product shots together make up half of all survivors. They are the least "creative" ads in the set: a number, a result, the product on a clean background. They last because they answer the question the buyer is already asking in the feed, what is this and what do I get, before the thumb moves on. See the feature callout examples for what this looks like at its best: one feature, stated as the whole ad.
2. Comparison ads live longest
Only 10% of survivors are comparisons, but their median lifespan of 96 days is the highest of any mechanism. A buyer weighing up a purchase is already comparing. The ad that does the comparison for them, on the brand’s chosen ground, removes a step. Pet food and beauty devices lean on it hardest in our data. Comparison ad examples.
3. Lifestyle imagery barely survives
Nine ads out of 276. Mood shots of people enjoying a product are what most brand teams make first and what the market keeps last. When lifestyle does survive, it is usually carrying a claim or a feature in the copy, so the image is decoration rather than the argument.
What about UGC?
UGC is 9% of the long-running statics, behind every mechanism above it in the table. That is not a case against UGC; it is a case against UGC as a format to tick off. The UGC ads that survive open on a specific problem and get to the product quickly. Social proof shows the same pattern from the other side: only 14 ads, but some of the longest-lived in the library, close to three years, and every one leads with a hard number such as a review count or customers served.
It changes by category
- Beauty and skincare (83 classified): feature callouts first, then before-and-afters. 16 of the library’s 22 before-and-after survivors are beauty brands. Buyers want proof they can see.
- Supplements and wellness (108): feature callouts, big claims, then offers and UGC. The product’s result is invisible for weeks, so the ad carries the trust.
- Food and beverage (49): feature callouts, offers and product shots. Repeat purchase pays for the first-order discount.
- Pet: nine of twelve survivors are comparisons against what the cat or dog eats now.
Category pages show the survivors in context: beauty, supplements, food and drink, home.
What to do with this
- Brief the mechanism, not the format. "Make a UGC video" is not a brief. "Compare our protein per pound to the supermarket brand, in the first second" is. The ad creative brief guide has a template.
- Start with the plain version. Before testing anything clever, test the feature callout and the big claim. They are the base rates.
- Look at your own category’s survivors first. The mix above is an average across 40 brands. Your two closest competitors’ longest-running ads are a sharper brief than any average.
- Judge the winner on profit. Survival tells you what works for someone else’s margin. Your own test needs POAS, not CTR. See how to test ad creative on Meta.
If you are generating creative with AI, all of this matters more, not less. Variants are now cheap; knowing which mechanism to make variants of is the scarce part. We cover that in AI ad creative.
Limits of this data
- Survival is a proxy. A long-running ad might be a small retargeting ad. Ads that run across many placements and in several versions are the stronger signal.
- Statics only. This classification covers static ads. Video is 62% of what the brands run and is analysed separately on each brand page.
- One label per ad. Most good ads stack mechanisms, a claim with a number with a review count. We label the one doing the main work.
- DTC brands, mostly UK and US. B2B and high-ticket categories behave differently.
You are welcome to cite these figures; please link to this page as the source.
Frequently asked
What type of ad creative works best on Facebook and Instagram?
In our data, plain feature callouts are the most common mechanism among long-running DTC ads (24% of 276 classified), followed by big claims (14%) and product hero shots (13%). Comparison ads are less common (10%) but have the longest median lifespan, 96 days. Lifestyle imagery accounts for only 3% of survivors.
How do you know an ad is working if you can't see its results?
Days in market. The Meta Ad Library shows when each ad started running but not its spend or results. An advertiser re-approves a live ad with their own budget every day, so an ad still running after 100 days has almost certainly paid for itself.
Do UGC ads outperform branded creative?
Not on survival. UGC makes up 9% of the long-running ads we classified, fewer than feature callouts, big claims, product shots, offers or comparisons. The UGC ads that last open with a specific problem and reach the product quickly; polish is not what keeps them live.
Are video ads better than static ads?
Video makes up 62% of the 6,146 ads we track, so brands are making far more of it. This study classifies statics only. Both formats survive; the mechanism matters more than the format, and the same mechanisms show up in long-running videos.
How often is this data updated?
The library is re-checked daily for whether each ad is still live, and the example pages for each mechanism are rebuilt from it. The figures in this guide are as of 23 October 2026.