Audience overlap is the proportion of people who belong to two or more of your Meta audiences at once. If 40% of the people in a “home improvement” interest audience are also in a lookalike audience of your past customers, those two audiences overlap by 40%. Meta has offered an Audience Overlap tool in Ads Manager to measure this, which is where the term usually comes from.
How audience overlap works
Every audience is a set of Meta accounts. Interest audiences, lookalikes, website visitors and customer lists are built in different ways, but they all draw from the same population. Two audiences that look different on paper can contain many of the same people, especially in a defined area such as Greater Manchester or a 10-mile radius around a shop.
Where your account still shows it, the Audience Overlap tool sits in the Audiences section of Ads Manager. You select up to five saved or custom audiences and ask for the overlap, and it reports how many people each pair shares as a number and a percentage. It works on audiences you have saved, so ad sets built with one-off targeting need saving as audiences first. The related overlap rate is simply that shared percentage.
Overlap is not a problem in itself. It becomes one when overlapping audiences are used in different ad sets with the same goal, because those ad sets then compete for the same people. That competition shows up as auction overlap, where Meta enters only one of your ad sets into each auction and the others underdeliver.
Why it matters
Overlap is one of the quieter reasons a UK small business account performs below its potential. A bathroom fitter in Nottingham might run one ad set for an interest audience, one for a lookalike and one for Page engagers, and believe they are testing three separate groups. If those audiences share most of their people, the test is not really a test, and each ad set’s results are shaped by the others.
Measuring overlap before launching also helps with frequency. Heavily overlapping ad sets can show the same person several different ads in a short period, which pushes up frequency and tires your audience faster.
Overlap also muddies reporting. When the same person could have been reached by any of three ad sets, the credit for their purchase lands on whichever one happened to win the auction, which says little about which audience is better.
Common mistakes
- Assuming different names mean different people. Interest audiences built around related topics overlap heavily.
- Checking overlap once. Website and engagement audiences change daily, so overlap does too.
- Splitting lookalike sizes into separate ad sets. A 1% lookalike sits entirely inside a 3% lookalike, unless the larger one is built as a 1% to 3% range.
- Treating overlap as the only cause of poor delivery. Weak creative and low budget per ad set matter as much.
How to act on it
Save the audiences used by your active ad sets and compare them with the overlap tool, or by reasoning about how they were built if the tool is not available to you. Where two ad sets with the same objective share a large part of their audience, merge them. Where you need them separate, such as prospecting and retargeting, use exclusions so each ad set keeps its own people.
For genuine audience tests, use Meta’s A/B test feature, which splits people into groups that do not overlap. Working through audience structure like this is part of my Facebook ads management service.