Genyad Labs · Research Brief

The Decay Curve: how fast a winning ad actually dies in 2026

We collected the frequency, CTR, hook-rate and lifespan figures scattered across platform write-ups and agency studies, checked each one against its source, and put them on one page. The consistent finding: a Meta ad that used to last about four weeks now fatigues in two to three.

Fig. 1: An illustrative CTR decay pattern on Meta, by week since launch
CTR index Week 1 = 100
wk2: −14%, still healthy wk3: −33%, fatigue confirmed wk4: −57% Week 1 Week 2 Week 3 Week 4 100 75 50 25
Mako Metrics' example pattern (CTR 2.1%, 1.8%, 1.4%, 0.9%), indexed. Illustrative, not a measured average. Source 06, see §5
Embed this chart on your site

Free to use in articles, decks and newsletters. Keep the source line.

Your site blocks iframes? Use the image instead:

01 / Headline numbers

Six figures worth knowing, and where each comes from

None of these are new data. They are published across platform write-ups and agency blogs. Each is quoted here as its source states it, metric included, with the source named.

2–3 wks
Time for a Meta ad to fatigue in 2026, down from about four weeks
Segwise, via Mako Metrics. AdLibrary sees the same on Reels-heavy placements
2.5
Weekly frequency on Meta prospecting above which performance starts to decline
AdLibrary. Past 4.0 it "falls off a cliff"
45%
Fall in conversion likelihood by the fourth exposure to the same creative
Meta's analysis of about 26,000 test cases (May 2023), via Darkroom
9 days
Median lifespan of a food & beverage ad, with CTR down 40%. The fastest vertical measured
Prestyj 2026 benchmarks, via Web Tonic
3–5×
Longer campaign life for brands running 15–50+ new variants a month, vs. refreshing quarterly
Prestyj 2026, via Web Tonic
31%
Lower CPA for brands testing 10+ creative concepts a month than for those testing fewer than 5
Motion Creative Benchmarks 2026, via Web Tonic

Two early-warning figures are worth adding. Mynt Agency treats a 15–20% CTR drop from peak within a 7-day window, with frequency rising at the same time, as the most reliable early signal of fatigue (via SHNO). And in the same 2023 Meta analysis, refreshing creative in ad sets that showed fatigue lifted conversion rates by about 8% on average (via Darkroom).

Embed these six numbers on your site

Free to use in articles, decks and newsletters. Keep the source line.

02 / Platform benchmarks

Fatigue looks different depending on where the ad runs

Medians from Benly's dataset of 12,347 video and static ads across 842 ad accounts, January to March 2026. Where Benly does not break a figure out for a placement, the cell says so.

PlacementMedian CTRMedian hook rate
Meta, video (all placements)1.62%28%
Meta Reels1.78%31%
Meta feednot broken out26%
TikTok in-feed video0.84%33%
YouTube in-stream0.42%22%

Frequency thresholds come from a different source, AdLibrary's guidance for 2026:

PlatformSafe weekly frequencyDecline starts above
Meta prospecting1.8–2.52.5 (cliff past 4.0)
TikTok2–33.5
YouTube TrueView2–33
LinkedIn (B2B prospecting)3–44

AdLibrary attributes the faster burn on Reels-heavy placements to Meta's Andromeda ranking, which it says weights creative-level features more aggressively than the system before it. That is AdLibrary's reading of the accounts it monitors, not a statement from Meta. Segwise, cited by Mako Metrics, points to Advantage+ concentrating delivery instead.

03 / Fatigue speed by vertical

Some categories burn through creative in days, not weeks

Median ad lifespan in ecommerce, with the CTR drop that marks fatigue in each category. Food & beverage and fashion sit at the fast end; home & garden lasts longest of the categories measured.

Food & beverage
9 days, −40%
Fashion
11 days, −35%
Beauty & skincare
14 days, −28%
Electronics
16 days, −31%
Health & supplements
18 days, −22%
Home & garden
21 days, −19%

Prestyj 2026 industry benchmarks, as reported by Web Tonic. Bar length shows speed of fatigue (shorter lifespan, longer bar).

No source here publishes a lifespan for B2B or SaaS ads. The nearest figures: Benly reports SaaS and B2B ads running 18% below the average hook rate and 22% below the average CTR, and AdLibrary puts the safe weekly frequency for LinkedIn B2B prospecting at 3 to 4.

04 / What this means for creative volume

Throughput, not talent, is the bottleneck

Two of the figures above make the same point from different directions. Brands running 15–50+ new variants a month extend campaign life three to five times compared with quarterly refreshers (Prestyj), and brands testing 10+ concepts a month see 31% lower CPA than those testing fewer than 5 (Motion).

Our reading, which is interpretation rather than a finding of any one source: the number of distinct concepts a team ships each month does more to keep an account ahead of fatigue than the quality of any single ad. Each additional variation has traditionally needed a new script, a new edit, a new voiceover and a new export per platform, which is why most teams produce fewer than the lifespans above call for.

This is the gap Genyad is built to close: turning existing footage into new scripts, voiceovers and platform exports without a reshoot, so keeping fresh variations in rotation is a production question, not a budget one. See how the workflow works →
05 / Methodology & sources

How this report was built

This is a synthesis, not a primary study. Genyad Labs collected publicly reported benchmarks and agency research on ad fatigue published up to October 2026. Every figure is quoted as its source states it, including which metric it measures (CTR, conversion likelihood, lifespan), and attributed to whoever produced it. Where the page we read cites someone else's data, we name both.

Have a dataset that confirms, refines or contradicts these figures? We will update this report and credit the source. Email contact@genyad.com.

06 / Corrections

What changed on 17 October 2026

We rechecked every figure in the August version against the page it cited. Several did not match, and we have corrected them:

Your creative is decaying right now. It's just a matter of how fast.

Genyad turns the footage you've already shot into the variation volume these lifespans call for.

See how it works