Spot creative fatigue on Meta and TikTok, and know what to make next.

How do I check ad fatigue in Meta Ads?

You can check for ad fatigue in Meta Ads with five steps in Ads Manager, no third-party tool needed.

1. Group by creative, not by ad. A creative that graduated from a testing campaign into a scaling one tires across both. Group by ad creative in Ads Reporting, or filter Ads Manager to every ad using the same post.
2. Compare the creative with its own past. Set the last 7 days against the previous 7, or against the creative's first 30 days. An account average mixes new and old creatives and hides the slope.
3. Read four signals together. Frequency, link CTR, CPM and cost per result. One metric moving alone usually has another cause; the table below shows which pattern points where.
4. Check Meta's own delivery status. Ads Manager can mark an ad Creative limited when its cost per result is higher than your past ads achieved, and Creative fatigue when it is at least twice as high (Meta Business Help Center, checked October 2026; also covered by Jon Loomer, 2025). It arrives late, because cost per result has to double first.
5. Judge prospecting and retargeting separately. Retargeting audiences are small, so frequency climbs faster there and a weekly frequency of 5 can be normal.

You'll find a frequency of 2.5 quoted as Meta's fatigue threshold on many blogs. We couldn't find it in Meta's documentation, and Meta's own fatigue statuses are based on cost per result. Jon Loomer, who has tracked Meta's ad tools for years, puts it plainly: "If results aren't tanking, I don't care what your frequency is."

How many impressions do you need before you call fatigue?

A fatigue call on a single creative often rests on too little data. Statistically, a week-on-week CTR comparison is a test between two proportions, and its sample size follows a standard formula. At 95% confidence and 80% power, the impressions needed in each window are n = 7.85 × [p1(1−p1) + p2(1−p2)] ÷ (p1 − p2)², where p1 is the CTR in the earlier window and p2 the CTR you want to be able to detect. At a 1% CTR, confirming a 20% drop takes about 35,000 impressions in each week. The table is our calculation.

Conversion rate is harder still: confirming a 20% fall from a 2% conversion rate needs about 17,300 clicks per window, roughly 1.7 million impressions a week at a 1% CTR. Few single creatives get there, so CPA alone flags fatigue too late to act on.

Three checks before you call fatigue on a creative:

• Set a minimum volume. Don't call fatigue on a creative below the impressions its CTR and the drop size need. Below it, stretch the window to 14 days.
• Aggregate by concept or hook. Five cutdowns of one UGC concept share much of their fatigue, so add them up. In the one field study that separates the two, the concept-level part of wear-out decayed about twice as fast per exposure as the part specific to each execution (Braun and Moe, Marketing Science 2013, one advertiser's banner campaign).
• Require two signals. Frequency rising with CTR falling, or CTR falling with cost per result drifting up. A single metric crossing a line is noise more often than fatigue.

How is TikTok ad fatigue different?

TikTok ad fatigue shows up first in attention, because viewers decide within a second or two whether to keep watching. TikTok Ads Manager reports what you need (TikTok Business Help Center): 2-second video views, 6-second video views, average play time per video view, and views at 25%, 50%, 75% and 100% of the video. Hook rate and hold rate are practitioners' names for ratios built from them.

Attention metrics also solve the volume problem. A hook rate sits far above a CTR, so the same formula needs far fewer impressions: a fall from 25% to 20% can be confirmed with about 1,100 impressions per window, against 35,000 for a 20% CTR drop at 1%. On TikTok you can call a tired hook within days.

TikTok publishes no refresh cadence. Its official guidance is 3 to 5 unique creatives per ad group, and Ads Manager suggests adding more when an ad group has fewer than three (TikTok Business Help Center). The "refresh every 5 to 7 days" rule you'll find online comes from vendor blogs. Measure your own instead: for your last ten TikTok winners, count the days until the hook rate fell 20% below its first week, and use the median as your refresh interval.

What does creative fatigue look like on app and gaming networks?

On app install campaigns the fatigue metric is IPM, installs per 1,000 impressions. Read each creative's IPM against its own first weeks on AppLovin, Unity and the app campaigns you run on Meta and TikTok. The volume rule from above applies unchanged: an IPM of 5 is an install rate of 0.5% per impression, so confirming a 20% IPM drop needs about 70,000 impressions in each window.

App campaigns add two traps. A high-IPM creative that promises a different game from the one people install brings cheap installs that churn, so check day-7 retention or ROAS by creative before scaling it (Lancaric on creative frameworks). Fatigue can also arrive from outside your account. When a format becomes the category default, players have seen the idea in other games' ads before they meet yours, and copycat creatives tend to stop scaling within a week (AppAgent).

Nupact reads IPM decay on gaming and app networks, with AppLovin and Unity connected for data and AppsFlyer or Adjust for installs. It acts on Meta and TikTok app campaigns. On AppLovin and Unity it flags and recommends, and your team makes the change.

Why did the creative win, and what should you make next?

Frequency caps and audience exclusions keep the worst cases away, and a tired ad gets paused eventually. Choosing its replacement depends on knowing why it won in the first place.

Most of the money sits in a few winners. In a study of about 1 million creatives on the Gunosy ad network in Japan, the 18.6% that ran for seven days or more generated 81.8% of sales (Kitada, Iyatomi and Seki, 2022). Knowing when a winner is wearing out, and what it did that the others didn't, pays more than catching losers early.

To answer why a creative won, tag every asset by hook (problem shot, founder to camera, unboxing), format, angle, offer and market, then read results by tag. One tagged winner tells you little. Ten winners and forty losers tagged the same way show which hooks and offers keep converting in which market, and that is the next brief.

Plan production from the retirement rate. If two winners retire each week and one concept in five becomes a winner, you need ten new concepts a week to keep the same number of winners live. Recuts of a winning hook are cheaper and extend a winner's life, so plan for those too.

A winner in one European market is often the cheapest next creative for another, and moving it means remapping the pixel, the product catalogue, the landing URL and the copy. Check all four before launch, since a Dutch ad that still sends people to the German landing page spends the NL budget on a page they can't read.

Which approach to creative fatigue fits your team?

Each of the common approaches is enough for some teams.

What should a creative fatigue tool do?

Use this list when you compare creative fatigue detection tools, including ours.

• Track each creative and concept across every ad set, campaign and platform it runs in.
• Compare against the creative's own history, with a minimum-volume rule before it alerts.
• Tell creative fatigue apart from audience saturation, because the fixes differ.
• Read each platform's own signals: frequency and CTR on Meta, 2-second and 6-second view rates and watch time on TikTok, IPM on app networks.
• Tag assets by hook, format, angle and offer, so the next brief starts from what won.
• Say recut or replace, and draft the change for someone to approve.
• Log every rotation with its approver, and let you revert it.

What else do people ask about creative fatigue?