Home · Guides

Guides

Creative fatigue: how often do you need new ads?

On Meta, a winning ad lasts two to six weeks at meaningful spend. To stay ahead of fatigue you need new concepts every week, not just new colours on the same ad. At €10k a month that is roughly 12 new creatives; at €50k it is 40 to 50; above €100k, 70 or more.

The creative studio in Ambition AI: AI produces the first version, a senior designer finishes it, every ad passes the claim check.
The creative studio in Ambition AI: AI produces the first version, a senior designer finishes it, every ad passes the claim check.

What creative fatigue actually is

Fatigue is not the ad getting worse; it is the audience getting used to it. The first people who see a new ad are the ones most likely to respond. As frequency rises, the remaining audience is less interested, so CTR drops, CPM rises and CPA follows. On Meta this shows up in the numbers before anyone notices it in the feed.

The practical definition we use: an ad is fatigued when its CPA is 25% above its own first two weeks, at similar spend, for three days in a row. Not when someone in the team is bored of it.

How many new creatives you need per month

New creatives per month by Meta spend, based on our accounts
Monthly ad spendNew conceptsTotal creatives (with variants)Typical lifetime of a winner
Up to €10k4–6124–6 weeks
€10k–€50k8–1230–483–5 weeks
€50k–€150k12–2050–722–4 weeks
Above €150k20+80+1–3 weeks

New creatives per month by Meta spend, based on our accounts.

Concepts beat variants

Ten variants of one concept are not ten creatives; they are one test with ten colours. Fatigue happens at the concept level: once the audience has seen the before/after angle, a new background does not reset the clock. Real refresh means a new angle, a new avatar or a new format.

Our rule for a monthly batch: at least four to eight distinct concepts, each with two to three hook variants. The concepts come from research (what wins in your account, what competitors run, what customers say), a coverage grid shows which avatar and funnel stage has no ad yet, and a creative strategist picks what gets made.

Where AI helps and where it does not

AI has made the first version of a static or a video cut cheap and fast. It has not made the decision cheap: which angle, for whom, why now. Brands that let AI generate 200 variants a month usually see fatigue faster, because the variants are all the same concept in different clothes. Brands that use AI for speed and a senior designer for the last 20% get ads that look like the brand and convert.

The workflow that works: research, coverage gaps, eight concepts from AI, a strategist picks three, the studio makes them in two to three variants, everything passes the claim check, winners get more variants next month.

How to read the signals early

Watch three numbers per concept, not per ad: frequency, CTR relative to its first week, and CPA relative to the account average. Frequency above 3 on prospecting is the first warning. CTR down 30% from week one is the second. CPA above your break-even is the moment to pause, not the moment to start briefing. Brief two weeks before you need the ad.

Frequently asked questions

Short answers to the questions we get on this topic.

Does creative fatigue also happen on Google?

Less, because search ads respond to intent rather than interrupt it. It does happen on Performance Max and Demand Gen assets, which behave like Meta. Refresh those on the same rhythm.

Can I just duplicate a winning ad to reset it?

Duplicating resets the learning, not the audience. The same people see the same ad. Change the concept or the audience, not the ad ID.

How long should I let a new ad run before judging it?

Until it has around 50 conversions or seven days, whichever comes first. Judging on two days and 300 euros is how good concepts get killed.

Related guides

More from the Ambition team.