Meta creative fatigue shows up as a weekly frequency above 2.5 on prospecting with CTR falling underneath it, and it arrives faster than most refresh schedules assume. Meta ads now fatigue in about two to three weeks, down from about four, and top-performing ads lose roughly 38 percent of their effectiveness after five weeks running unchanged. The fix is a rotation sized to that curve, which for a normal Meta account means four to six new concepts every fortnight on feed and closer to weekly on Reels, not one refresh per quarter.
All of those figures come from our 2026 ad fatigue benchmark, a synthesis of published platform and agency figures rather than our own measurement.
What does creative fatigue look like on Meta specifically?
Meta gives you a frequency number, which most platforms do not, and that makes the diagnosis easier here than anywhere else. Frequency is average impressions per person over your reporting window, so it is only meaningful with a window attached. A frequency of 2.5 over seven days and 2.5 over ninety days are different situations entirely, and reading a lifetime frequency as a fatigue signal is the most common mistake we see.
Set your window to the last seven days, split by ad set, and read frequency against CTR in the same view. The pattern that confirms fatigue rather than a bad creative is both lines moving: frequency up, CTR down, over consecutive weeks. A creative that launched at a low CTR was never good. A creative whose CTR has halved while frequency climbed is fatigued, and those need different responses.
Meta's 2023 analysis of about 26,000 test cases, quoted in our benchmark, is the number to plan around: about 45 percent lower conversion likelihood by the fourth exposure to the same creative. That is the mechanism underneath the frequency band. It is not that the platform punishes you, it is that the fourth viewing of the same opening two seconds is not a new event for the person watching.
The frequency bands for Meta
| Weekly frequency | What is happening | What to do |
|---|---|---|
| Under 1.5 | Audience has room, creative is still meeting new people | Nothing. Scale budget if the numbers support it |
| 1.5 to 2.5 | Normal working range for Meta prospecting | Start the next batch of concepts now, before you need them |
| 2.5 to 3.0 | Decline begins. Our benchmark puts the Meta prospecting ceiling at 2.5 | Rotate in replacements this week |
| Above 3.0 | Fatigue has arrived. Expect CTR well below launch | Swap the creative set, not the budget |
| Above 3.0 on retargeting | Often normal, because the pool is small by design | Judge on cost per result instead, cap frequency if available |
Two qualifications. Retargeting frequency runs higher than prospecting frequency by construction, so applying the 2.5 band to a 20,000-person remarketing pool will have you refreshing creative that is performing perfectly well. And frequency is a symptom, not the disease: broadening an audience lowers frequency without making the creative any newer, which buys a few weeks and does nothing for the underlying decay.
For a specific read on how long your current set has left, our ad fatigue calculator takes your frequency, your CTR trend and your vertical and works out where you are on the curve.
Does Reels fatigue faster than the feed?
The benchmark medians for the two Meta placements are closer than most people expect.
| Placement | Median video CTR | Median hook rate | Frequency ceiling |
|---|---|---|---|
| Meta Reels | 1.78 percent | 31 percent | 2.5 |
| Meta feed | Not published separately (1.62 percent across all Meta placements) | 26 percent | 2.5 |
| TikTok, for contrast | 0.84 percent | 33 percent | About 3.5 |
| YouTube in-stream | 0.42 percent | 22 percent | 3 per week |
Similar hook rates, the same frequency ceiling, and a Reels CTR slightly above the Meta median across all placements. So the fatigue tolerance is not really different between the two Meta surfaces. What differs is the speed you reach the ceiling. Reels inventory is cheaper and impression velocity is higher, so the same budget accumulates frequency faster there, and a creative that would take a fortnight to burn in the feed can burn in a week on Reels.
That gives a straightforward cadence rule: refresh Reels weekly and the feed fortnightly, then adjust for your vertical rather than treating either number as fixed.
Vertical changes the timing more than placement does. Median ad lifespan, with the CTR drop that marks fatigue, from the benchmark:
- Food and beverage: 9 days, CTR down 40 percent, the fastest measured
- Fashion: 11 days, down 35 percent
- Beauty and skincare: 14 days, down 28 percent
- Electronics: 16 days, down 31 percent
- Health and supplements: 18 days, down 22 percent
- Home and garden: 21 days, down 19 percent
The benchmark has no lifespan figure for B2B or SaaS. A food brand running Reels is on a nine-day clock. A home and garden brand running feed has three weeks, which is why the same fatigue advice reads as paranoid to one team and complacent to the other. Take your vertical's number, shorten it if you are on Reels, and set your production cadence to beat it rather than to react to it.
The refresh volume that follows from the curve
This is the part that decides whether any of the above is actionable. If Meta ads now fatigue in about two to three weeks, and in our experience a typical active campaign needs 8 to 20 live variations in rotation, then you are replacing a meaningful share of your library every fortnight, permanently.
Doing the arithmetic for a mid-sized Meta account:
- 12 live variations across feed and Reels
- Three-week useful life each, so roughly a third retiring every week
- That is four new concepts a week, or 15 to 20 a month
- The benchmark's own finding is that brands shipping 15 to 50 variants a month see 3 to 5 times longer campaign lifespan than quarterly refreshers
The report's operational conclusion is blunt: throughput, not talent, is the bottleneck. The number of unique concepts a team ships per month predicts campaign longevity better than the quality of any single ad. That runs against how most creative teams are organised, which is around perfecting one hero asset per campaign, and it is the single most useful thing in the data.
What does not count toward that number: the same video in four aspect ratios, the same video with a new caption font, the same video trimmed by two seconds. Fatigue is a function of what the viewer recognises, and they recognise the opening shot. A refresh that keeps the first two seconds is not a refresh.
What to change when you rotate
Ranked by how much of the fatigue they actually reset, in our experience:
- New opening two seconds, new footage. The full reset. Different shot, different first line, different setting.
- A different argument. Move from demo to testimonial, or from problem-solution to comparison. This resets fatigue for the concept and often reaches a different segment.
- New hook on the same body. Cheap, fast, and worth perhaps half the lift of a full reset. Good for stretching a winner an extra week.
- New voiceover or a different narrator. Genuinely helps on sound-on placements, does very little on silent autoplay in the feed.
- New captions or new music. Marginal. Do it while doing something on this list, not instead.
- Raising the budget on a fatigued winner. Actively harmful. It buys more frequency on the creative that is already exhausted.
Genyad is our product, and this rotation problem is what it exists for. It takes a library of footage you already own, tags every clip, and builds each variation as a fresh script, shot selection, voiceover and caption set rather than a re-cut of the same timeline, which is what makes item one on that list repeatable rather than a shoot request. One variation is one credit, the free plan includes five with no card, and every variation exports at 9:16, 4:5, 1:1 and 16:9 without a watermark. It will not publish to Meta for you, it has no predicted performance score to tell you which variation will survive longest, and it cannot invent footage you never shot, so a library with three usable clips will produce variations that look like three usable clips rearranged.
Frequently asked questions
What frequency is too high on Meta ads?
On prospecting, a weekly frequency above 2.5 is where decline begins according to our benchmark report, and above 3.0 the fatigue is already showing in CTR. On retargeting, higher frequency is normal because the pool is small, so judge those campaigns on cost per result rather than on the frequency number itself.
How often should I refresh Meta ad creative?
Weekly on Reels and fortnightly on the feed for most accounts, adjusted for vertical: median ad lifespan is about 9 days in food and beverage, 11 in fashion, 14 in beauty and skincare, 16 in electronics, 18 in health and supplements and 21 in home and garden. There is no published figure for B2B or SaaS. Set the cadence to arrive before that date, not after it.
How long does a Meta ad creative last?
Meta ads now fatigue in about two to three weeks, down from about four. In Mako Metrics' illustrative example CTR is 14 percent down in week two and 33 percent down in week three, and top-performing ads lose roughly 38 percent of their effectiveness after five weeks running unchanged. Two to three weeks is the number to plan production around.
Will changing the aspect ratio or the caption fix fatigue?
No. Fatigue is driven by recognition, and people recognise the opening shot, so a re-crop or a new caption font is the same ad to the person scrolling. A real refresh changes the first two seconds: new footage, new opening line, or a different argument altogether.
How many creative variations should a Meta account have live?
Between 8 and 20 live variations per active campaign, replaced at roughly a third per week given a three-week useful life. That works out to 15 to 20 new concepts a month for a mid-sized account, and the benchmark links shipping 15 to 50 a month to 3 to 5 times longer campaign lifespan than refreshing quarterly.