How often do advertisers actually refresh their creative?

40 advertisers measured on 18 August 2026. The median has been running its current ads for 40 days. The range is 1 day to 277.

The distribution

≤19d

Top quartile — refreshes fastest. MANSCAPED (1 day), Altra Running (5), Jolie (6).

40d

Median — the midpoint of the whole sample.

≥58d

Bottom quartile — slowest to change. Casely (277 days), Vuori (173), DERMAFLASH (167).

The finding that surprised us

How many ads an advertiser runs tells you almost nothing about how fresh they are. Across these 40 brands the correlation between active-ad count and median ad age is r = 0.07 — statistically, no relationship at all.

Two brands make it concrete. Casely runs 726 active ads with a median age of 277 days: the second-largest library in the sample, and essentially none of it is current work. MANSCAPED runs 34 ads with a median age of one day. Any comparison based on volume gets these two exactly backwards.

That matters because volume is the number vendors quote. It is also the number that accumulates on its own if nobody cleans up. Age is harder to flatter, because it decays — stop producing and your median climbs by one every day.

How to read a number here

A long-running ad is not automatically neglect. Advertisers kill losers quickly, so an ad still live after 200 days is probably one that works. And a short median is not automatically good — rapid churn can mean high testing velocity, or that nothing is landing.

What the number reliably captures is activity. A set where everything is months old and nothing new has appeared is the clearest available signal that nobody is currently working the account.

All 40, fastest to slowest

Advertiser	Active ads	Median age	Over 60d
MANSCAPED	34	1	0%
Altra Running	115	5	0%
Jolie	22	6	0%
Bombas	29	13	14%
Vivrelle	47	13	3%
Parachute Home	188	13	0%
Graza	246	13	7%
Warby Parker	79	14	0%
Shop FlavCity	210	15	0%
Sono Bello	238	16	0%
Solo Stove	29	19	10%
Flintts	202	23	13%
Alpha Lion	288	27	27%
Quince	4,009	28	7%
Halfday	59	29	0%
Beis	775	30	30%
BenchMade Modern	129	30	0%
Arrae	581	33	0%
Hunt A Killer	14	34	0%
Mad Rabbit	861	40	43%
Nutrisense	50	41	0%
Brooklinen	224	42	23%
Rugiet	570	44	43%
Dr. Squatch	496	44	20%
Avantera	80	48	40%
Sijo	53	49	0%
Olipop	24	50	42%
Caraway	249	50	37%
AG1	389	51	47%
Cotopaxi	109	54	43%
HexClad	273	58	43%
StopBox USA	400	58	43%
Liquid Death	274	69	67%
Tumble	324	82	77%
Silly George	97	97	96%
Chubbies	122	107	83%
feejays	13	118	62%
DERMAFLASH	76	167	100%
Vuori	876	173	97%
Casely	726	277	100%

What this is not

It is not a grade of any agency. These are advertisers, and a brand’s advertising may involve several agencies and an in-house team. Nothing here establishes who made which ad, and we will not publish a number that implies it does.

Method: Meta Ad Library, advertiser-scoped by page ID, US, active ads only, measured 2026-08-18. Per brand we read the ~30 ads on the first result page and take the median age, so per-brand figures are indicative rather than exact. Known limits: sort parameters are ignored by the endpoint so the sample cannot be reshuffled; responses intermittently return empty and must be retried or brands silently vanish; and brands with generic names cannot be isolated at all, which biases the sample toward distinctive names. This is one snapshot — a second measurement is scheduled to establish whether individual brands hold their position or move week to week. How we vet.