How to identify creative fatigue in Meta Ads
Creative fatigue is frequency rising and CTR falling, together, sustained over 14 days or more. Both signals, in that direction, over that long. High frequency on its own is not fatigue, and a single week of declining CTR is not either. This matters because fatigue is the most over-diagnosed condition in paid social — it is the explanation people reach for first, and at least half the time the real cause is delivery, tracking or the landing page.
To confirm creative fatigue, check:
- Frequency trend — is it rising, not just high
- CTR trend over the same window — is it falling
- Whether both have moved together for 14+ days
- Whether CPA followed, and with what lag
- Creative age and how long it has carried budget
- Audience size — a small pool explains high frequency without fatigue
- Whether new creative in the same ad set is also declining
Fatigue, and the four things that look like it
The signature matters because the fixes are completely different. New creative will not solve a delivery problem.
| What you see | What it probably is | What to do |
|---|---|---|
| Frequency rising AND CTR falling, 14+ days | Creative fatigue | New creative — not a budget change |
| Frequency high but flat, CTR stable | A small audience pool | Normal for retargeting. Leave it alone |
| CTR falling, frequency flat | A new weak creative dragging the average | Find the ad, not the format |
| CTR falling across new and old creative alike | Delivery shifting to worse segments | Check breakdowns and budget changes |
| CTR stable, CVR falling | Landing page, offer or stock | Not a creative problem at all |
| Everything flat, conversions down | Tracking | Check pixel and CAPI event volume |
Confirming it properly
Step 1
Pull daily data, not a period total
Fatigue is a trend, so a single aggregate cannot show it. Pull daily ad-level data across at least 21 days so you can see the shape and find the date the lines crossed. The inflection date is what lets you connect the creative decline to whatever happened in the account that week.
Rate metrics cannot be averaged across rows. Pull CTR and frequency at the level you intend to read them rather than reconstructing them from daily figures — reach and frequency in particular cannot be recomputed from lower-grain data at all.
Step 2
Check both signals, in the same window
Frequency rising and CTR falling, over 14 or more days. One without the other is not fatigue, and this is where most false diagnoses happen. A retargeting ad set against a small pool runs at frequency 6 permanently and converts fine; that is the pool size, not decay.
Step 3
Isolate it to the creative, not the ad set
Check whether every ad in the ad set is declining or just the older ones. If new creative entering rotation performs normally while the old ads decay, that is fatigue. If new creative arrives and immediately underperforms too, the problem is delivery or audience — adding more creative will not fix it.
Step 4
Split the video funnel to find where it decayed
For video, fatigue shows up in different places depending on what wore out. Hook rate is 3-second views over video plays; hold rate is thruplays over 3-second views. A declining hook rate means the opening has stopped stopping people. A stable hook with declining hold means they still stop, but the body no longer holds them.
The thresholds we use: hook rate at or above 40% is good, under 25% poor; CTR at or above 1.25% is good, under 0.65% poor. Hold rate is judged against the account's own 90-day average rather than a fixed figure. These are our working numbers, not an industry standard.
Step 5
Follow the chain to CPA
Fatigue has an order: frequency climbs, CTR falls, CPM effectively rises because you are paying for impressions that do not click, and CPA follows last. Establishing that sequence with dates is what separates a diagnosis from a correlation — "CTR started falling on the 9th, CPA began rising on the 13th" is an argument; "both look bad this month" is not.
Step 6
Rule out the alternatives before you accept it
Check the account change log over the same window. A budget increase pushing delivery into worse segments produces falling CTR without any creative decay. A new campaign competing for the same audience raises frequency. An attribution change moves CPA without touching creative at all.
A worked example
Illustrative example. The numbers below are made up to show the method. They are not a client account and not a benchmark.
One ad, tracked across three weeks. This is what the signature looks like when it is real.
| Window | Frequency | CTR | Hook rate | CPA |
|---|---|---|---|---|
| Days 1–7 | 1.8 | 1.52% | 44% | $38 |
| Days 8–14 | 2.6 | 1.21% | 38% | $41 |
| Days 15–21 | 3.9 | 0.84% | 29% | $57 |
Both signals move together, in the right directions, across three weeks. Frequency more than doubles while CTR nearly halves. That is the signature.
The hook rate decline localises it. Falling from 44% to 29% says the opening stopped working — people are no longer stopping to watch. That points at a hook rebuild rather than a whole new concept.
CPA follows with a lag, which is what you would expect. It barely moves in week two and then jumps 39% in week three. Waiting for CPA to tell you about fatigue means acting two weeks late.
The check that makes it conclusive: if a newer ad in the same ad set held its CTR across the same period, this is creative fatigue. If the newer ad declined too, something changed in delivery and new creative would not have helped.
What not to conclude too quickly
Fatigue is the default explanation for almost any decline, which makes it the one worth questioning hardest:
- High frequency alone is not fatigue. Small retargeting pools run high frequency indefinitely and convert well.
- One week of declining CTR is not a trend. Hold the 14-day bar.
- A falling account-level CTR may be one new weak ad dragging an average, not a format decaying.
- Delivery changes mimic fatigue precisely. Increasing budget pushes delivery into worse segments and CTR falls with no creative involvement.
- Seasonality moves CTR without anything being wrong.
- If CVR fell but CTR held, the creative is doing its job — look at the landing page, offer or stock.
- Below 3–5 conversions a week, the CPA swing you are reacting to is noise.
How to automate fatigue detection
Catching fatigue early means tracking two metrics per ad, daily, across every ad carrying meaningful spend, and spotting the moment their trends diverge — before CPA moves, because by then you are two weeks late. Nobody does that by hand reliably.
GoMarble tracks the signals continuously against the account's own baseline and surfaces the divergence when it appears, rather than when someone next opens the creative report.
The shape of it
- Connect the ad account
- Creative metrics are tracked daily per ad
- Frequency and CTR trends are evaluated together, not in isolation
- The alert names the ad and the date the trends diverged
- Recommended next creative directions come attached
The point of catching it early is that you build the replacement while the winner is still winning.
Creative fatigue checklist
- ☐ Pull daily ad-level data across 21+ days
- ☐ Check frequency is rising, not merely high
- ☐ Check CTR is falling across the same window
- ☐ Confirm both have moved together for 14+ days
- ☐ Check whether newer creative in the same ad set is also declining
- ☐ Split hook rate from hold rate on video
- ☐ Judge hold rate against the account's 90-day average
- ☐ Establish the sequence and dates through to CPA
- ☐ Check audience size — a small pool explains high frequency
- ☐ Check the account change log for budget or targeting changes
- ☐ Rule out landing page and tracking if CVR moved but CTR did not
Questions people actually ask
What frequency is too high on Meta Ads?
How long does creative last on Meta?
Is creative fatigue the same as audience saturation?
Should I pause a fatiguing ad immediately?
Can I prevent creative fatigue?
Where this method comes from
The fatigue signature, the thresholds and the funnel split on this page come from the diagnostic signals GoMarble's agents apply to live Meta accounts — specifically that fatigue requires frequency rising and CTR falling together over 14 or more days, that high frequency alone does not qualify, and the hook and hold rate formulas with their benchmarks. Those benchmark figures are ours rather than industry standards.
Related
Catch fatigue before CPA moves.
Connect Meta and GoMarble tracks frequency and CTR together — alerting when they diverge, not after.