Frequency calculator

Impressions is the number of times the ad was served over the period, as the platform counts it.

Reach is the number of distinct people it was served to. This is the field that decides whether the answer means anything, because reach is a claim about identity. Platforms resolve it differently — by account, by device, by household, by probabilistic matching — and a person with two devices may be one reach or two depending on whose report you are reading.

Target frequency is optional and deliberately not supplied. There is no correct frequency in general, only a correct one for a given ask, audience and flight length.

A frequency of four is an abstraction. The impressions behind it are a media plan.

At a reach of 210,000 and a target of three, the campaign would serve 630,000 impressions. It actually served 840,000, so the target implies 210,000 fewer — a quarter of the delivery, and at any given cost per thousand a specific amount of money. Stated that way, the decision about whether to widen the audience or hold the frequency is a budget conversation rather than a stylistic one.

The same line works upward. A campaign at 1.2 against a target of three needs to serve exactly two and a half times the impressions it has, which is either a much larger budget or a much smaller audience, and putting the number on the page usually settles which.

Exposure is not evenly distributed. Heavy users of a platform are available to be served far more often than occasional ones, so the same campaign that reports an average of four may be showing the ad twice to most people and fifteen times to a minority. Those two groups have entirely different experiences of the campaign, and only one of them is at risk of being irritated by it.

This is why an average frequency should be read as a trend rather than as a fact about a person. A campaign whose average has climbed from two to five over three weeks is being served repeatedly into a pool it has exhausted, whatever the distribution looks like. Where a platform exposes the distribution itself, use it; the share of reach above some exposure count is a much better description of fatigue risk than any mean.

Frequency also cannot be added across platforms. A person reached on two of them is two reach records and no report knows they are the same person, so a total frequency across an account is a number with no referent.

Almost nobody cares about frequency itself. What they care about is one of two things, and frequency is a cheap indicator for both.

Fatigue. As a creative is repeated to the same people, the rate at which they act on it falls. The click-through rate usually falls first and the cost per result follows, which makes a rising frequency with a falling rate the earliest reliable warning available in a paid social account. The cause is rarely the number itself: it is that the same execution has stopped being new.

Audience exhaustion. A campaign that cannot find new people to reach will keep spending by serving more impressions to the people it has already found. That is a targeting and budget-fit problem wearing a frequency costume, and the fix is a wider definition or a smaller budget rather than a cap.

There is a case in the other direction too. A frequency that never gets above one on a considered purchase means nobody is being given a second chance to notice, and campaigns that spread a small budget across an enormous audience routinely fail this way while reporting excellent reach.

Find your own turning point rather than importing one. Plot the cost per result against frequency across several flights in your own account; the point at which cost begins to climb is the frequency at which your creative, your audience and your offer stop compounding. That figure is worth writing down and revisiting when the creative changes, because a new execution moves it.

Then choose the lever that matches the cause. Widening the audience lowers frequency and changes who is reached. Extending the flight lowers the rate of exposure without changing the total. Adding creative does not lower frequency at all, and it is often the right answer anyway, because it changes what is being repeated. A frequency cap is the mechanical option: it works, it costs delivery, and it is best set as a boundary the campaign should not cross rather than as a dial to tune weekly.

Finally, read frequency beside reach rather than alone. Frequency rising while reach is flat is exhaustion. Frequency rising while reach grows is a campaign doing its job at a larger scale, and cutting it in that state buys nothing.

And treat a frequency target as a property of a flight rather than of an account. The right number for a two-week sale push is not the right number for an always-on prospecting campaign, and an account-wide cap applied to both will under-serve the first and over-serve the second.