Creative refresh cadence gets treated as a single question with a calendar answer, usually every two weeks. That rule has no source, and following it produces two opposite mistakes at once: retiring adverts that are still earning, and keeping ones that stopped.
The fix is to split it into two questions. Produce on a schedule. Replace on a measurement.
This is for you if somebody has told you your creative is stale because of how long it has been running.
Why one creative refresh cadence cannot cover both jobs
Production and replacement are different activities with different constraints.
Production is a capacity question. How many new adverts can your business actually make, given who makes them and what else they do. That has a rhythm and a schedule suits it.
Replacement is a performance question. Is this specific advert still producing results at an acceptable cost. That has no rhythm, because adverts do not decline on a timetable.
Merging them into one rule means the timetable overrides the measurement, which is exactly backwards. The calendar knows nothing about your account.
So keep a production schedule and a replacement trigger, and let them run independently.
The production side, set by spend
The most useful published guidance sets production rates by account size, and the bands are lower than most advice implies.
Below roughly ten thousand a month, about one new creative monthly. From ten to twenty five thousand, three to four. Between twenty five and fifty thousand, roughly one weekly. Above that, up to a hundred thousand, two to four weekly.
The reason spend drives this is exposure. A larger budget reaches more people more often, so it consumes creative faster and needs replacing sooner.
It also funds it. An account at fifty thousand a month can support producing weekly. One at five thousand cannot, and pretending otherwise produces variants shipped as concepts.
Find your band and treat it as the floor for production planning. That is a resourcing conversation rather than a marketing one, and it is worth having explicitly.
The replacement side, set by cost
Replace when the number says so, and there is exactly one defined threshold available.
Meta labels a creative as showing Creative Fatigue when cost per result reaches at least twice its past level, with a separate Creative Limited label when the cost is higher but has not doubled.
That is the only threshold with an author. Everything else circulating, including the fortnightly rule and every frequency cap, was invented downstream.
Use Limited as your early warning and Fatigue as the outer marker. In practice I start planning a replacement when something shows Limited and act when the cost keeps climbing, rather than waiting for a clean doubling.
Nothing about elapsed time enters that decision, which is the point.
Why the fortnightly rule fails both ways
It is worth being specific about the damage, because the rule sounds harmless.
You retire creatives that are still working. An advert producing results at a stable cost gets replaced because two weeks passed, and you have thrown away accumulated performance and pushed the ad set back into unstable delivery for nothing.
And you keep creatives that stopped working. Cost climbs on day four and everybody waits until day fourteen, because that is when the review happens.
Both are expensive and they happen in the same account in the same month. That is the signature of a rule that has no relationship to the thing it governs.
The measurement trigger avoids both, and it costs no more effort once it is set up.
Creative refresh cadence should rotate partially, never wholesale
A practical detail that gets this wrong even when the trigger is right.
Replacing an entire ad set at once throws it back into unstable delivery and destroys the history that made your comparisons readable. You lose the working creative alongside the failing one, and the set spends a week finding its footing again.
Replace what has genuinely declined. Leave the rest running. New creatives enter alongside the survivors and the set keeps its accumulated learning.
This also gives you a cleaner read on the new material, because it is competing against a known quantity rather than against four other unknowns.
The exception is a wholesale change of direction, where the offer or the audience has genuinely moved. Then a clean start is correct, and it should be a deliberate decision rather than a side effect of a rotation schedule.
What to do when production cannot match the band
Most accounts are below their band, and there are two honest responses.
Tighten the audience. A smaller audience is exposed more slowly to your existing creatives, so the same material lasts longer. You give up reach and buy time, which is the right trade when production is the binding constraint.
Prioritise distinct angles over finished executions. Three rough adverts making genuinely different arguments will outlast six polished versions of one idea, because the audience experiences the six as repetition.
What does not work is running fewer creatives to make them last. Low creative count is the most common cause of fatigue arriving early, so that response accelerates the problem it is trying to solve.
Being below your band is not a failure. It is a constraint on how fast the account can scale, and naming it clearly is better than pretending a fortnightly rule solves it.
Where creative refresh cadence stops applying
Search is a different case and the same logic does not transfer.
Responsive search assets are combined by the system rather than shown as complete adverts, so no single advert accumulates exposure the way an image does. Text also does not wear out visually the way a photograph does after the fortieth viewing.
Refresh there is driven by message testing rather than by exposure. You change the copy because you want to test a different claim, not because the old one is tired.
So a cadence built for paid social should not be applied wholesale to search, and plenty of agencies do exactly that because it makes the reporting look consistent.
What to change this week
Three steps.
Find your production band from your monthly spend and compare it against what you actually produce. That gap is the honest picture of your creative capacity.
Then find the Creative Limited and Creative Fatigue labels in Ads Manager and use them as your replacement trigger, because they are the only defined thresholds available.
Then drop any calendar rule you currently follow, and check whether anything currently running would have been retired by it despite performing.
Creative fatigue has one real definition covers the threshold in detail, ad frequency thresholds are made up covers the exposure metric people usually reach for instead, fund the ad set and read what the algorithm chose is the position underneath, and the weekly rhythm I run is the lead generation version. Content refresh works on the same logic organically, which my SEO work treats the same way. To set a cadence that fits your actual capacity, book a teardown.