Why do my Meta ads die so fast?
Because the pocket the ad speaks to is exhausted, not because the creative went stale. If every ad in the account is built on the same desire, belief system and avatar, they all draw from one pool of buyers, so each batch fatigues faster than the last — the pool is smaller every time. Ads built on genuinely different positioning angles draw from separate pools and last longer.
The shape of the drop tells you the cause
A short, sharp spike followed by a fast decline is not a learning-phase problem and it is usually not an algorithm change. It is what it looks like when an ad finds the most responsive people in its pocket within days, converts them, and then has nowhere fresh to go.
Watch the two numbers together. CPM rising while CTR falls on the same audience is depletion. The auction is getting more expensive at the same time the message is landing with fewer people — which is what happens when you have moved past everyone the message was built for.
Compare that to a genuine creative miss, where CTR is weak from the start rather than strong-then-collapsing. Depletion has a peak. A bad ad never has one.
Why each batch dies faster than the last
This is the part that makes the pattern unmistakable.
The first ad into a fresh pocket reaches the whole pocket, including the people most ready to buy. It runs well and runs for a while. The second ad into the same pocket reaches what is left. The third reaches what is left of that.
So the spikes get shorter and smaller, and the gap between launching and dying keeps compressing. Teams read this as "creative is getting harder" and respond by producing more, faster — which accelerates the same curve.
Nothing about the creative is degrading. The pool is.
"But all our ads are different"
Most brands believe they practise creative diversity. Different hooks, different formats, different visuals, a real spread of concepts.
Lay them out side by side and read them for who they are addressed to, and the picture usually changes. Same underlying desire. Same assumed beliefs. Same person implied by the casting and the copy. Twelve ads, one pocket.
Format variety is not audience variety. You can change the hook, the edit, the colour and the talent and still be talking to exactly the same person in a slightly different voice.
The test is simple: for each ad, write down the desire it names, the belief it assumes and the person it pictures. If those three columns are the same all the way down, you do not have twelve ads. You have one ad, twelve times, and it will fatigue like one ad.
What actually extends lifespan
Ads last longer when the pocket underneath them is fresh, so lifespan is decided before anything is designed.
Test angles with statics first. Not because video does not work — it does. Because statics strip away the noise. No retention editing, no trending audio, no creator charisma. Just the positioning and the copy. If an angle works as a static, it usually explodes as video. If it only works as video, you do not have a winning angle, you have a winning performance — and that is why those winners cannot be replicated.
Run enough variation per angle to learn something. Five to ten ads for each new angle. Fewer than that and you cannot tell whether the angle failed or that particular execution did.
Give each angle its own ad set. One ABO campaign, one ad set per angle, ads grouped underneath. Nothing clever. This forces Meta to give each angle a real test instead of funnelling budget to whichever is cheapest on day one.
Leave it alone for 3–7 days. Most decisions are clear by day five. Budget each ad set at roughly 2–4× your target CPA — at a €50 tCPA that is about €150/day.
Refresh before the drop, not after. Once you know roughly how long an angle lasts in your account, the next one should be in test before the current one peaks. Structured refresh ahead of the decline is the whole difference between managing fatigue and reacting to it.