Meta frequency cap on a small prospecting audience: set one or let it run?

Small Meta accounts keep asking whether to put a frequency cap on prospecting ad sets. Take a local service targeting a 40k audience in one metro. Frequency hits 4 or 5 within a week, CTR slips, and someone in the comments says cap it at 2. On an audience that small, a hard cap mostly stalls delivery, so you end up paying more for the same reach. I'd rather watch CPA and frequency together and rotate creatives before frequency gets high. Where a cap does make sense is retargeting, since high frequency there just annoys people who already said no. Has anyone seen a cap actually lower CPA on prospecting, or does it always come down to how fast the creative rotates?

Rotation usually matters more on a 40k audience, agreed, but a binding cap mostly makes you underspend rather than pay more per reach. The budget just sits there and the algorithm gets fewer chances to learn. So I'd test it by looking at pacing first. Say a dental clinic sets a 7-day cap of 2 and by day four has spent only 60% of its daily budget. That gap is the real cost of the cap, in lost conversions rather than higher CPM. If spend stays near budget and CPA drops, keep the cap. If not, take it off and fix the creative.

I don't think a binding cap mostly just leaves budget unspent. On a 40k audience, once the people who've hit the cap drop out, the system keeps spending on whoever's left, and that's usually the colder end of the audience. So a cap can hurt CPA even while pacing looks healthy. Take a dental clinic capping at 2 with spend landing at 95% of budget: CPA can still creep up over the week, because the remaining impressions go to people who've barely seen the brand.

That's why pacing alone is a weak check for your test. Look at CPA by frequency bucket, or compare the audience mix (age, placement, time of day) before and after the cap goes on. If spend holds but the mix gets colder, the cap is costing you even though the pacing chart looks fine.

The cold-mix effect is real on a small audience, but I wouldn't read it off frequency buckets. Frequency climbs for everyone across the week, so a higher bucket is mostly just a later day, and creative fatigue hits those same days. The CPA creep you're seeing could be fatigue rather than the cap.

A cleaner test is two ad sets with identical creative and budget, one capped and one not, run over the same two weeks. Compare CPA and the share of impressions going to people at low frequency. If the capped set holds its CPA while the uncapped one drifts up, the cap is earning its keep.