Facebook Ads Audience Size Calculator
How much of an audience a budget can reach and at what frequency, with audience overlap priced, and the platform estimate called what it is.
Audience estimate 1,400,000
what it is accounts the platform models as matching, not people
Budget $6,000.00 over 14 days
a day $428.57
CPM $8.40
Impressions the budget buys 714,286
At a frequency of 2.5
people reachable 285,714
share of the audience 20.4%
cost a person reached $0.0210
If the budget covered the whole audience
frequency you would get 0.51
budget for the target frequency $29,400.00
which is more than you are spending, so the audience is larger than the budget can cover
the consequence the platform will reach the cheapest slice of the audience rather than all of it
What each audience size would give
350,000 frequency 2.04 if fully covered
700,000 frequency 1.02 if fully covered
1,400,000 frequency 0.51 if fully covered ← yours
2,800,000 frequency 0.26 if fully covered
5,600,000 frequency 0.13 if fully covered
Overlap with the other audience
the other audience 900,000
overlap 35% of the smaller, about 315,000 accounts
combined unique 1,985,000
duplicated share 15.9%
what it costs above about a fifth, two ad sets are bidding against each other often enough to raise the price of reaching people you had already won
The audience figure is an estimate of accounts the platform models as
matching, not a count of people. For some targeting combinations these
estimates have exceeded the census population of the country in
question, because they include duplicate and inactive accounts. Use it
as an order of magnitude and never as a population.
This budget buys about 714,286 impressions, which covers 285,714 people
at a frequency of 2.5: 20.4% of the audience. Budget, audience size and
frequency are one decision rather than three, and fixing any two
determines the third.
The audience is larger than the budget can cover, so the platform will
deliver to the cheapest slice of it. That slice is not a random sample:
it is the people who are cheapest to reach, which is not the same as the
people most likely to buy.
Overlapping ad sets bid against each other in the same auction, which
raises what you pay to reach somebody you were going to reach anyway.
Meta publishes an audience overlap tool for exactly this, and the fix is
consolidation rather than more exclusions.
A narrow audience is not the same as a good audience. Broad targeting
with strong creative outperforms narrow targeting on most modern ad
platforms, because the delivery system finds the responsive people
faster than a marketer can define them.
Lookalike percentages are about similarity, not size: a 1 percent
lookalike is the closest 1 percent of the country to your source, and it
is only as good as the source. A lookalike built from all purchasers is
usually worse than one built from the best purchasers.
Reach saturates rather than stopping. As a campaign exhausts the
responsive part of an audience the cost per result rises steadily, which
looks like fatigue and is often exhaustion, and the two are fixed
differently.
Output is valid and updates as you type.
Fix the highlighted fields to update the output.
Budget, audience size and frequency are one decision rather than three. Fix any two and the third is determined, which is why an audience of 1.4 million and a budget of $6,000 cannot also be a frequency of 2.5 across everybody: the money reaches 285,714 people, or 20 percent of the audience.
What happens to the other 80 percent matters more than it sounds. The platform does not sample the audience randomly. It delivers to the cheapest slice of it, and cheap to reach is not the same as likely to buy.
The audience figure itself deserves less trust than it gets. It is an estimate of accounts the platform models as matching, not a count of people, and for some targeting combinations these estimates have exceeded the census population of the country.
How to use
- Put in the audience size the platform shows, the budget and the CPM you expect.
- Set the frequency you are aiming for. Two to three is the usual range for a campaign of a few weeks.
- Add a second ad set’s audience and its overlap to see how much of your spend is bidding against itself.
Example
$6,000 over 14 days at an $8.40 CPM, against an audience of 1.4 million:
Impressions the budget buys 714,286
At a frequency of 2.5
people reachable 285,714
share of the audience 20.4%
cost a person reached $0.0210
If the budget covered the whole audience
frequency you would get 0.51
budget for the target frequency $29,400.00
which is more than you are spending, so the audience is larger than the budget can cover
the consequence the platform will reach the cheapest slice of the audience rather than all of it
What each audience size would give
350,000 frequency 2.04 if fully covered
700,000 frequency 1.02 if fully covered
1,400,000 frequency 0.51 if fully covered ← yours
2,800,000 frequency 0.26 if fully covered
Overlap with the other audience
overlap 35% of the smaller, about 315,000 accounts
combined unique 1,985,000
duplicated share 15.9%Narrowing to 350,000 would buy a frequency of 2 across the whole audience on the same money. Whether that is better depends on whether the smaller audience is actually more likely to buy, which the platform’s estimate cannot tell you.
Pitfalls
Potential reach counts accounts, not people. Duplicate accounts, inactive accounts and modelled matches are all in there. Use it as an order of magnitude and never as a population.
An audience larger than the budget is not free optionality. It hands the delivery system the choice of whom to skip, and it optimises for cost rather than for value.
Overlapping ad sets bid against each other. Above roughly a fifth duplication, you are paying more to reach people you had already won. The fix is consolidating ad sets, not adding more exclusions.
Narrow is not the same as good. Broad targeting with strong creative beats narrow targeting on modern delivery systems, because the algorithm finds responsive people faster than a marketer can define them.
Lookalike percentages are about similarity, not size. A 1 percent lookalike is the closest 1 percent of a country to your source list, and it is only as good as that source. Built from all purchasers it is usually worse than one built from the best purchasers.
Reach saturates rather than stopping. As a campaign exhausts the responsive part of an audience, cost per result climbs steadily. That looks like creative fatigue and is often audience exhaustion, and the two are fixed differently.
The CPM you enter is a forecast. It moves with season, placement, competition and creative, and it is highest in the fourth quarter. Run the numbers at a range rather than a point.
Compatibility
Arithmetic in the browser: nothing is uploaded and nothing is stored.
Impressions come from the budget and the CPM, people reached from impressions divided by the target frequency, and the coverage share from that against the audience. The sensitivity table holds the budget and CPM fixed while moving the audience, which is the comparison that shows the three-way trade directly.
The overlap is expressed as a share of the smaller audience, which is how Meta’s own audience overlap tool reports it, and the combined unique figure is the two audiences less the duplicated accounts. The test suite asserts the reading flips when a budget becomes large enough to cover its audience, since that is the point where the advice changes from “narrow” to “widen”.
Coverage is capped at 100 percent, so a budget larger than its audience reports full coverage rather than an impossible share.