You're Probably Showing Your Ads to the Same Person Too Many Times

The Number That Made the Audit Awkward

A national QSR franchise changed their paid social agency. Part of the handover was an audit of what the previous setup had been doing.

The finding: their retargeting campaigns were showing ads to the same people 20 to 30 times per week.

Not 20 to 30 times over a campaign flight. Per week.

The marketing team had been running retargeting against a relatively small audience - people who had visited the website or interacted with content - without adequate frequency caps. As the audience pool stayed roughly stable in size, the algorithm kept serving impressions to the same users to hit spend targets.

At the 25th impression, the incremental value of that ad is effectively zero. The person has either converted, actively decided not to, or learned to ignore your creative entirely. Spending money on the 25th impression doesn’t drive a transaction. It just costs money.

Why Over-Frequency Happens

Paid social platforms are designed to spend budgets. Their algorithms will find a way to deliver impressions and hits against their targets regardless of whether the audience is large enough to absorb that volume without over-serving.

When a retargeting audience is too small for the budget allocated to it, the algorithm recycles. The same users keep getting served because they’re the eligible audience. The platform doesn’t flag this as a problem - from its perspective, it’s delivering the impressions you’re paying for.

Frequency caps exist as a manual control, but many accounts either don’t set them, set them too loosely, or set them at the campaign level rather than the user level. A cap of “3 per day” at the campaign level is meaningless if the user is in five overlapping retargeting audiences that are all serving simultaneously.

The result: a significant portion of retargeting spend goes to people who have already received multiple exposures and haven’t converted - people for whom additional impressions are not going to change the outcome.

The Incrementality Connection

This is an incrementality problem. The question isn’t whether impressions are being delivered - they are. The question is whether those impressions are causing transactions that wouldn’t have happened otherwise.

For the first few impressions, there’s a reasonable chance of incrementality: the person is reminded of the brand, considers a visit, and follows through. But by impression 20, almost all of the incremental effect has been extracted. The people who were going to be moved by the ad have been moved. The people who weren’t going to convert still aren’t going to.

Marketing mix models capture this through adstock and saturation functions: the incremental value of each additional impression in a given period declines as frequency increases. At extreme frequency levels, the model shows near-zero (and sometimes negative) incremental contribution.

Negative contribution sounds counterintuitive, but it can happen. Over-served ads can create ad fatigue that actually reduces purchase intent. Someone who’s seen the same creative 25 times may develop a subconscious negative association with the brand.

The Small Audience Problem

Over-frequency is most common when retargeting budgets exceed the carrying capacity of the audience.

A retargeting pool is defined by the behaviour you’re capturing: website visitors in the last 7 days, video viewers, engagement with posts, email list custom audiences. These pools are typically a fraction of the brand’s total reach - often 1-5% of the people seeing top-of-funnel activity.

When media plans allocate a meaningful budget to retargeting without adjusting for audience size, the frequency per person goes up fast. A $50,000/month retargeting budget against an audience of 20,000 people means $2.50 per person per month. Against 200,000 people, that’s $0.25 per person. Against 2,000 people, that’s $25 per person - extremely high frequency territory.

The fix is mechanical: build audiences large enough to absorb the budget at reasonable frequency, set user-level frequency caps rather than campaign-level caps, and actively monitor impression-per-unique metrics in the platform.

What Changed

For the franchise in question, the new agency restructured the paid social targeting around two changes.

First, they broadened the retargeting criteria to increase pool size - capturing a wider set of intent signals rather than only website visitors, which was the smallest and most saturated audience.

Second, they introduced hard frequency caps at the user level and shifted budget from the over-served retargeting pool toward prospecting activity targeting new audiences.

The result was the same or better performance from a lower cost base - not because the creative was better, but because the impressions were being served to people who hadn’t already been exposed 20 times.

The MMM measurement confirmed this: when frequency normalised, the incremental coefficient for Facebook improved, meaning each impression was generating more incremental revenue than it had been when frequency was at peak over-saturation.

A Practical Frequency Audit

If you run paid social retargeting, the following checks are worth running monthly:

Frequency report. Most platforms show average frequency per user. A weekly average above 5-7 in a retargeting campaign is a signal to investigate. Above 10 is a problem.

Audience size vs. budget ratio. Divide your weekly retargeting spend by your audience size. If the per-person weekly spend is high relative to your average order value or consideration cycle length, you’re likely over-serving.

Incremental lift test. Run a holdout test on your retargeting audience. Serve ads to 80% and hold out 20%. Measure conversion rate difference. If the holdout group converts at nearly the same rate as the served group, your retargeting spend is going to people who would have converted anyway.

Creative refresh. Frequency fatigue accelerates when the creative is static. If you’re running the same three creatives for 90 days, even reasonable frequency levels can produce fatigue effects. Rotate more aggressively.

Over-frequency is one of the most common and most correctable paid social inefficiencies. It’s also one of the most invisible, because the platform reports show impressions and click-through rates without flagging the per-person saturation underneath.

Key Takeaways

• 20-30 impressions per week to the same user has near-zero incremental value - most of that spend is wasted
• Over-frequency happens when retargeting budgets exceed the carrying capacity of the audience pool
• Platforms won’t flag this - they’re designed to deliver impressions, not optimise for incrementality
• User-level frequency caps, audience size expansion, and holdout tests are the main controls
• MMM confirms the effect: incremental contribution per impression improves when frequency normalises
• Broadening retargeting criteria (more intent signals) and shifting budget toward prospecting is often more efficient than continuing to serve a saturated audience