A tradesperson running ads within a 20 kilometer radius of his workshop often sees the same thing after three or four weeks: a cost per lead that was creeping up slowly suddenly takes off, even though nothing changed in the campaign settings. Same budget, same targeting, same creative. That last part is exactly the problem: the creative is the same, and the same people have seen it too many times.
This has a name, ad fatigue, and a measure, frequency. In a small local area it shows up much faster than in a national campaign, for a purely arithmetic reason worth understanding before looking for explanations elsewhere. This piece is a companion to our creative testing method for Meta Ads: that article covers how to test new angles, this one covers when the old ones stop working.
Why a small area burns out faster than a national market
Meta serves your ads to an audience pool that depends on your geographic targeting, age range, and any interest layers you add. A business selling sunrooms, kitchens, or heat pump installations typically targets a 15 to 40 kilometer radius around its service area, sometimes one or two departments. That audience might represent a few tens of thousands of relevant people, against several million for a national brand. How wide to make that pool is the question behind Advantage+ versus manual local targeting.
The math is mechanical. For a given daily budget, Meta spreads a fixed number of impressions across a pool of people. The smaller that pool, the faster each individual sees the same ad again. A national audience of 5 million people absorbs a €30/day budget for months without much movement in individual frequency. A local audience of 40,000 people can hit an average frequency of 3 or 4 in a few weeks with the exact same budget.
Three factors make this worse for a local, high-ticket service business:
- Targeting is deliberately narrow. You don't want to pay to reach someone 200 km away who can never become a client, so you tighten the radius, which mechanically shrinks the available pool.
- The decision cycle is long. A sunroom or renovation project gets decided over several weeks, sometimes months. The same person stays in the active audience much longer than an impulse buyer, and therefore sees the ad more times before either acting or tuning it out.
- Too few creatives are running at once. A single ad, or two very similar variants, exhaust a small audience pool far faster than a pool of five or six distinct angles.
The symptom isn't abnormal. It's the direct consequence of the otherwise sound choice to target a narrow zone and a qualified audience instead of running broad and expensive for nothing.
Reading frequency: the number to check every week
Meta shows frequency in Ads Manager at the ad set or campaign level, by adding the corresponding column to your dashboard. It reports the average number of times someone in the audience has seen the ad over the selected period.
Some practical benchmarks for a local lead generation campaign:
- Frequency below 2 over 7 days: healthy, the audience is refreshing properly, nothing to change.
- Frequency between 2 and 3.5 over 7 days: watch zone. Not a problem yet, but the moment to prepare the next creative rather than wait for a drop.
- Frequency above 3.5 to 4 over 7 days: fatigue is generally already there or very close. Click-through rate typically starts declining and cost per result starts climbing.
These thresholds aren't universal laws, they vary by industry and by the creative itself, but they give a reliable order of magnitude for making a call without waiting for the bank account to speak for you.
The right habit is to read frequency alongside cost per lead over the same period, never one without the other. Frequency rising without any cost per lead degradation isn't urgent: the audience still tolerates the repetition. Frequency rising at the same time cost per lead climbs is the clear signal to act. When cost per lead keeps climbing, the levers that lower cost per lead show what to change besides the creative.
The signal that doesn't lie: the cost-per-lead curve over time
Frequency alone is never enough to decide. The decisive signal remains how cost per lead moves week over week, compared on an identical basis (same budget, same targeting, same campaign schedule). Three shapes come up most often:
- Stable plateau. Cost per lead oscillates within a reasonable range with no clear trend. Nothing to do, the campaign is running in its normal efficiency zone.
- Gradual drift. Cost per lead rises 10 to 20% over two to three consecutive weeks, alongside frequency crossing the thresholds above. This is the typical fatigue case: the creative keeps running, but it convinces fewer and fewer people in an audience that has already seen it a lot.
- Sharp drop. Cost per lead doubles or triples within a few days. This is rarely fatigue alone, it often signals a combination, advanced fatigue plus an external event (school holidays, news cycle grabbing attention, a competitor just entering the same area). In this case, refreshing creative alone won't always fix everything.
The point of caution for a local service business, where monthly lead volume stays modest compared to national e-commerce, is not reacting on too little data. A three or four day increase on ten total leads isn't a trend, it's statistical noise. You need at least one full week, ideally a comparison across two consecutive weeks, before concluding there's real fatigue rather than normal fluctuation. Major edits can also restart the learning phase, which is why the Meta Ads learning phase explained simply is worth knowing before you touch a running ad set.
When to refresh creatives: the trigger, not the calendar
The classic mistake is scheduling creative refreshes on a calendar, "every month" for example, without looking at the numbers. That's like swapping a car part that still works while leaving the worn one in place. Refresh should be triggered by signals, not by date.
Three concrete triggers, in the order they should raise attention:
- Frequency above 3.5 to 4 over 7 days, even if cost per lead is still holding: time to prepare the next creative, not necessarily to cut immediately.
- Cost per lead rising two weeks in a row, correlated with frequency: refresh to handle within the week.
- Click-through rate dropping more than 30% versus its average over the first four weeks: an early signal, often visible before cost per lead even moves noticeably.
Refreshing doesn't mean rebuilding everything. The most effective method is keeping the angle that works (the promise, the proof, the format that performed best) and changing what's visually recognizable: the opening hook, the first frame of the video, the layout of the image. For the testing method itself, angle by angle and format by format, see our dedicated article: Meta Ads Creative Testing: the Method to Scale.
A reasonable refresh cadence in a local area
For a local audience pool, with a daily budget of €15 to €40, a refresh cadence that holds up over time looks like this:
- Two to three creatives running at all times, never just one. This spreads exposure and mechanically slows fatigue on each one.
- A new creative introduced every three to five weeks, ahead of the signal rather than in pure reaction, as soon as frequency enters the watch zone.
- Rotation, not wholesale replacement. Introduce the new creative alongside the existing ones, watch it for a week, then cut the one showing the most fatigue rather than changing everything at once and losing the campaign's learning history.
This cadence isn't set in stone. An industry with a short decision cycle, or a product that lends itself to a lot of visual variation (carpentry, kitchens) can tolerate more frequent refreshes. An industry where each creative is expensive to produce (video with an on-site crew) can space them out more, provided frequency is watched more closely to avoid being caught off guard.
The underlying thread stays the same: in a small geographic area, ad fatigue isn't a one-time anomaly to fix, it's a recurring phenomenon to monitor continuously. A business that checks its frequency and cost per lead every week anticipates the refresh. A business that only looks at these numbers when leads drop suffers through it, and pays more to catch up.