Ad frequency in Meta Ads for gyms: how many times is too many for a local audience

Everything you've read about frequency in Meta Ads was written for e-commerce with audiences in the millions. You run a gym in a town of 40,000 people with a 3 km radius, and your real audience is 8,000–15,000 people. At that scale, frequency isn't a metric to glance at occasionally, it's the hourglass for your campaign. It spikes in two weeks, not two months, and most of the "Facebook just doesn't work for me anymore" complaints we hear are exactly this: the same ad hammering the same neighbourhood until the neighbourhood learned to tune it out.

The math nobody does

Frequency isn't some algorithmic mystery. It's a division:

impressions per person = (budget × days ÷ CPM × 1,000) ÷ audience size

With numbers from a typical case: €20/day on an audience of 12,000 people with a CPM of €9. That's roughly 2,200 daily impressions, around 15,500 a week. Spread across 12,000 people: theoretical weekly frequency of 1.3. In practice, Meta doesn't distribute evenly, it concentrates impressions on those most likely to convert, so your highest-affinity core (maybe 3,000–4,000 people) has been seeing the ad 3–4 times a week from day one.

By week three, the cumulative campaign frequency in that core exceeds 10. They've seen your offer ten times. The ones who were going to click already did. The rest you're paying Meta to remind that you exist, which they already know.

Run your own numbers before reading on: monthly budget, audience size (shown in the ad set), average CPM. If the result gives you more than 4–5 weekly impressions per person in cold prospecting, your problem isn't the creative or the bid. You're pouring too much water into a small glass.

What frequency is healthy by campaign type

There's no universal magic number because context changes what "too much" means. Someone who already visited your website tolerates (and needs) more impressions than a stranger.

Measure frequency in 7-day windows, not cumulative campaign totals: a three-month campaign with a total frequency of 9 can be perfectly healthy (0.75/week) or severely burned out (concentrated in the last month). The default frequency column in Ads Manager is misleading; add the time breakdown.

The signs of saturation and how to tell them apart from creative fatigue

Audience saturation has three symptoms that arrive in order:

1. CTR drops week over week. The people who were going to engage already have; the rest have developed banner blindness to your logo.
2. CPM rises. Meta charges more to place an ad people are ignoring, because the auction penalises low engagement. You pay more for impressions worth less: the deterioration compounds.
3. Comments shift in tone. "This ad again," the odd angry reaction, the occasional hide. When you get here, you're already two weeks late.

Now the distinction that matters: this looks a lot like creative fatigue, but it's not the same condition even if the remedy looks similar. Creative fatigue means your piece wore out: the audience still has room, but that specific video no longer stops the scroll, and a fresh creative resets the numbers. Audience saturation means the well ran dry: you've already reached almost everyone receptive, and a new creative gives a breather of days, not weeks. The diagnostic test is exactly that: put in a fresh creative. If CTR recovers and holds, it was fatigue. If it bounces and falls again within a week, it's saturation, and you need to touch the audience or the budget, not the design. Fatigue signals and their timeline have their own article; the rule here is: same apparent remedy, different condition, different dose.

The 4 valves to reduce pressure

When frequency spikes, you have four valves. In order from least to most drastic:

1. More creatives rotating

Almost always the first response. Five active creatives instead of two spread impressions: the same person sees your brand 6 times but with 4 different ads, and the hammering sensation drops even if the technical frequency is the same. Different angles also give each subgroup its hook: back pain, early mornings, community. Keep 3–5 pieces live and replace 1–2 per month. This is the cheap valve and you should keep it open all the time, not just in a crisis.

2. Expand the radius or temporarily remove exclusions

More well. Going from 3 to 5 km can double your audience, and in many cities 5 km is still a realistic member distance (the test: look at where your current members actually come from, it's usually surprising). You can also temporarily relax non-critical exclusions or open the age range by one bracket. Watch the tradeoff: every extra kilometre brings people with a lower probability of joining. Expand only as far as your real member map tells you, not until the frequency can breathe.

3. Lower the budget

Yes, lower. In a small audience there's a monthly demand ceiling: in a given neighbourhood, there are N people willing to consider a gym this month, and spending double doesn't manufacture more. If at €600/month you have frequency 6 in cold prospecting, the last €200 is buying repeated impressions on burned-out people. Drop it to €400, let frequency come back to 3, and CPL improves while spending less. Almost no owner finds this natural (lower budget, better results?), but in local it's often the right answer. Do it gradually, 20% less every few days, to avoid jolting the learning phase; the stable budget lever explains why sudden changes are expensive.

4. Burst campaigns

The drastic valve: 2 weeks on, 1 off. It gives the audience time to half-forget you and gives you a "seasons" structure that pairs well with deadline offers. The cost is real: each reactivation restarts part of the learning phase, and the first 2–3 days of each burst are erratic and more expensive. That's why it only makes sense in very small markets (audiences below ~10,000 where valves 1–3 aren't enough) or businesses with strong seasonality where you'd concentrate spend anyway. For the average gym, continuous and lower budget beats bursts: the learning phase tuition is paid once, not twelve times a year.

The special case of the small town

If your town has fewer than 25,000 inhabitants, design from the start assuming saturation isn't a risk but the default state. Your complete audience might be 5,000–8,000 people and you'll see all of them every month. The game changes there: the goal stops being "find new people" (there aren't any) and becomes "stay present without exhausting people when their moment arrives." Low, consistent budgets, aggressive creative rotation, plenty of supporting local organic content, and bursts aligned with peak periods (January, September, the pre-summer season if it applies to your niche). The monthly success metric isn't the week's CPL, it's the cost per new member over the quarter.

Your frequency review, in 5 weekly minutes

Add to your routine: frequency column with 7-day breakdown, CTR and CPM in the same view. Three weeks of trend (CTR falling + CPM rising + frequency >5 in cold) = open valves, starting with valve 1. And before increasing budget on any campaign, repeat the division from the top: if the audience isn't growing, extra budget becomes frequency, not leads. That back-of-envelope calculation saves more money than any dashboard. If you also have budget split between prospecting and retargeting with their separate frequency thresholds, each euro works in the healthy zone of its own table.

Creative rotation is the valve that feels like the most effort to keep ready, because it requires producing new pieces every month indefinitely. It's also the part that makes the most sense to automate: tools like Pilotium generate and rotate creatives from your real photos precisely because in local audiences frequency doesn't forgive. With or without automation, the discipline is the same: in local, the ad wears out before the budget does, and the winners are those who accept that from the moment they design the campaign.