Effective frequency: how many times must someone see your ad?

A lone commuter waiting on a London station platform beside two large backlit advertising panels
Photo: mattbuck (category), CC BY-SA 3.0

"How many times does someone need to see an ad before it works?" is the most reasonable question a client ever asks, and the honest answer is that it depends on four things — none of which are the platform you happen to be buying.

Where "three exposures" came from

The number most people half-remember is three. It traces back to Herbert Krugman's argument in the early 1970s that the first exposure to an advertisement answers "what is it?", the second "what of it?", and the third is a reminder — or the beginning of disengagement. It was a thoughtful psychological model, not a media-buying rule, and it was written for a world of three television channels.

What survived is the useful half: the early exposures do different jobs, and none of them is persuasion. Planners later reframed this as effective frequency — the minimum number of exposures within a purchase cycle at which advertising starts to do measurable work.

The four things that actually set the number

FamiliarA known brand needs fewer exposures than an unknown one
ClutteredThe noisier the category, the higher the floor
ConsideredLong, risky purchases need presence across the whole cycle
  • How well they already know you. Recognition is a head start. A brand nobody has heard of is paying for the first two impressions just to become legible.
  • How crowded your category is. Frequency is relative. In a loud market, being seen eight times may be less present than four times in a quiet one.
  • How complicated the decision is. A £4 impulse buy and a £40,000 contract do not need the same rhythm of contact.
  • How good the creative is. A striking, distinctive ad earns attention on the second exposure that a dull one will not earn on the ninth. Frequency cannot rescue work nobody looks at.
Recall against frequency A curve showing brand recall rising slowly for the first two impressions, climbing steeply between three and ten, then flattening after about fourteen impressions a month. 136 101420 Impressions per person, per month Brand recall wasted saturation builds here diminishing returns
The curve, not the threshold. The first two impressions buy recognition, the middle ones do the work, and the far end buys irritation.

Why one number is the wrong shape of answer

Frequency is not a threshold you cross; it is a curve you climb. The first exposures buy recognition, the middle ones build the association you actually want, and the far end buys nothing but annoyance. The planning question is therefore not "what is the magic number?" but "where does our curve flatten, and are we anywhere near it?"

In practice, for a cold audience in a competitive UK market, plan for the region of eight to twelve exposures a month and measure what happens. Below three you are typically buying nothing. Above roughly fifteen, in most categories, you are paying to irritate people who already know exactly who you are.

Average frequency lies

Every ad platform will happily report an average frequency of 4.2 and let you believe your audience saw you four times. They did not. A small, heavily-online slice saw you twenty times, a long tail saw you once, and the average describes nobody.

Two habits fix this:

  • Look at the distribution, not the mean. What proportion of the audience reached the level you planned for? That is the number worth reporting.
  • Cap the top. Frequency caps are not a way to save money; they are a way to move wasted impressions from people who are saturated to people who are not.
Frequency distribution versus average frequency A histogram showing most of the audience receiving one or two impressions while a small group receives more than twenty. The reported average of 4.2 describes almost nobody. 1234 56812 1620+ Impressions received by one person in a month Reported average: 4.2 a frequency almost nobody received 62% saw you once or twice heavy users, over-served Frequency caps do not save money. They move impressions from the right of this chart to the left.
Why average frequency flatters every report. The mean sits in a gap between the many who saw you twice and the few who saw you twenty times.

How to plan it without guessing

  1. Define the audience narrowly enough that your budget can reach all of it, not a random sample of it.
  2. Set a target frequency and work out what it costs before committing — audience × frequency ÷ 1,000 × CPM.
  3. Produce enough creative that the eighth impression is a different execution of the same idea.
  4. Report achieved frequency every week, against the plan, alongside cost per customer.

Do that and effective frequency stops being a trivia question and becomes what it should be: a budget line you can defend.


Want this done for your market?

Send us your audience and budget and we will run these numbers properly, then tell you what saturation would cost and whether it is worth doing.

Get your free saturation plan Try the calculator

Ready to be the only name they see?

Tell us your market and we will map exactly where your buyers spend their attention — and what it costs to own it.

Call us Get your plan