Ask a media buyer what's holding a campaign back and the answer is rarely targeting or budget. It's creative fatigue: the same handful of ads served long enough that the audience has stopped seeing them, and performance quietly decaying while the account looks unchanged on the surface.
The number, roughly
If you're spending seriously on Meta, the working minimum is 10 to 20 fresh assets a month. That's not a volume target for its own sake, it's the rate needed to keep testing new hooks and formats while a handful of previous winners are still running. Below that, you're not testing, you're just waiting for the current set to stop working.
The number moves with spend. A brand testing a few hundred dollars a day can get by on the lower end. A brand pushing five figures a day burns through creative fast enough that 20 a month is closer to a floor than a ceiling.
Format beats quantity
Ten variations of one format teach you less than one variation each of ten formats. Before/after, stat panel, review grid, versus, timeline, receipt: these are proven structures for a reason, and the job each month isn't inventing new formats from scratch, it's running your product through the formats that are already known to convert and finding which ones your specific audience responds to.
That reframes the volume question. You don't need 20 completely original ideas a month. You need your product run through a wide enough spread of known-working formats, with enough copy variations per format, that the data tells you something real about what's working.
What actually breaks this
Most brands don't fail at knowing this. They fail at production capacity. A media buyer spending their week in Canva instead of in the ad account is the actual bottleneck, and it shows up as a plateaued account long before anyone diagnoses "creative fatigue" as the cause.
The practical fix is separating creative production from media buying entirely, so the buyer's job stays optimisation and testing, not asset production. Whether that's an internal hire, a freelancer, or a production partner depends on the brand's stage, but the principle holds regardless: whoever is watching the account shouldn't also be the one making the assets, because the moment volume drops, so does the testing cadence, and the whole account slows down with it.
If your last three winning ads are more than six weeks old and you haven't replaced them, that's the actual signal to act on, not a calendar reminder to make more content.