Meta creative fatigue: swap ads on frequency or on a CTR drop?

Lots of threads rotate creatives on a fixed schedule, every two weeks or so, and I don't think the calendar is the right trigger. On a prospecting ad set spending a few hundred a day, frequency climbs slowly, and CTR can drop for reasons that have nothing to do with the creative. If the auction got pricier that week, the two signals get mixed up. A better approach is to watch CTR and CPM together against a baseline from the first week the ad ran. If CTR falls around 25% while CPM stays flat, the creative is tired. If CPM jumps, it's the auction, and a new asset won't fix that. Is anyone using frequency as a hard trigger? The threshold seems to vary a lot with audience size, so I'm curious how people set it.

Agreed on the CPM check, that's the part most people skip. On the threshold, I'd drop the raw frequency number entirely, because it only means something relative to audience size. Take a local dental clinic targeting 150k people in one city: a frequency of 3 there can mean most of the reachable audience has already seen the ad, while on a 20M broad audience a frequency of 6 might be nothing.

What I'd watch instead is the share of impressions going to people who've already seen the ad, which Meta shows in the reach and frequency breakdowns. Once you're past roughly half the audience being saturated, a falling CTR stops being a creative problem and becomes an audience-size problem. The fix then is widening the audience or adding a new one, not just swapping the image.

The denominator that matters is the audience after exclusions. Most clinics exclude past patients, website visitors and people who already opened the lead form, and that can shrink a 150k city audience to 60k or less. An ad set that looks fresh at 150k can be half saturated at 60k within two weeks, even while frequency looks harmless.

Meta's audience size estimate in Ads Manager is only a rough range, so treat it as a ballpark and watch how fast reach flattens instead. If unique reach stops growing while spend keeps going, you've hit the ceiling, and that's a stronger signal than any frequency threshold.

Placement matters too. Stories and Reels tend to burn through a small pool faster than Feed, so a CTR drop in one placement can be a placement problem, not a creative problem.

Exclusions shrinking the pool is only half of it, because the setting that quietly changes the math is Advantage+ audience expansion. If it's on, Meta can push past your targeting, so the real denominator is murkier than the post-exclusion number you're looking at. Turn it off on a test ad set and the reach curve will show you the true ceiling.