How to find winning Facebook ads (with Meta's own reach as proof)
Published 13 Nov 2026. Data read 2026-11-12. Source: Meta's Ad Library (EU transparency data), read by Poseidon, unless the text names another source.
To find winning Facebook ads, look in the Meta Ad Library for ads that are still running after three weeks or more, then check how many people each one reached, which Meta publishes for every ad shown in the EU. Days live alone isn't proof: when brands launched several videos with the same text on the same day, the best one reached a median 6.5 times more people than the weakest (81 tests, read 12 November 2026).
Step 1. Keep only ads that are still running after 21 days
A brand pays for every day its ad runs and sees its own sales, so an ad still live after three weeks is one it chose to keep. Poseidon uses that line, 21 days, to call an ad a winner.
How far past 21 days real winners go, across 5,293 e-commerce ads still running in the EU on 12 November 2026:
| Days live | |
|---|---|
| A quarter had run less than | 55 days |
| Median | 115 days |
| A quarter had run more than | 194 days |
| Still running after 90 days | 59.3% of winners |
Step 2. Check how many people it reached
For every ad shown in the EU, Meta's Ad Library lists how many people in the EU saw it, and splits them by country, age and gender. Open the ad's details in the Library to read it. This is the number that tells two long-running ads apart.
Example. A weight-loss brand in France ran the same text on three different videos. The best reached 527,300 people in the EU, the weakest 127,920, four times less (best · weakest). Same brand, same words. Only one video is the one worth copying.
Step 3. Look for the brand's own tests
Brands test their ads against each other: the same text on different videos, the same video with different words. When several versions sit side by side, the one with the most reach is the one Meta kept showing. Across 81 such tests launched the same day, from 46 shops, the best version reached a median 6.5 times more people than the weakest, and at least 3 times more in 80% of them. See the brands' own A/B tests.
Step 4. Read the first line
Once you know which ad won, read its hook, word for word. That's the part you can test in your own ad. See the hooks that survive 90+ days in weight-loss ads.
Doing it by hand, or asking Poseidon
By hand, the Ad Library shows one ad at a time, with today's reach and no history. Poseidon reads it every day for every winning e-commerce ad it follows and keeps the history, so you can ask "which videos won brands' own tests in weight loss in France?" and get the ads back with their reach. Free pages by niche and country: winning ads by niche.
How this was measured
- Source: Meta's start date and EU transparency data for each ad, read by Poseidon on 11 and 12 November 2026, snapshot frozen at 22:48 UTC on 12 November.
- Which ads: e-commerce shops on Poseidon's watchlist, ads still active on 12 November and live 21 days or more.
- EU reach is Meta's count of people in the EU who saw the ad at least once. It's the only number of people shown. It isn't sales.
- Tests: ads of one brand sharing the same text, with a different video, launched the same day; the outcome is measured on EU reach.
FAQ
What is a winning Facebook ad?
An ad the brand keeps paying for. Poseidon counts an ad as a winner once it has run 21 days or more and is still active, then ranks winners by the reach Meta publishes.
Is the Meta Ad Library free?
Yes. Anyone can search it without an account. Reach, age, gender and country are shown only for ads delivered in the EU.
Does reach mean sales?
No. Reach is how many people saw the ad. A brand that keeps a high-reach ad running for months is paying for it, which is the strongest public sign you can get, but it isn't a sales figure.
How long should a winning ad have run?
At least 21 days. Half of the 5,293 winners read on 12 November 2026 had run more than 115 days.
Ask Poseidon what's winning. askposeidon.app