TERM · MARKETING

Lookalike audience

What is a lookalike audience?

A lookalike audience is a group Meta builds by finding people who resemble a source list you supply — customers, site visitors, or people who watched a video. You choose how close the match is: 1% of a country’s users is the tightest and most similar, 10% the broadest. The output is only ever as good as the source list behind it.

Example

A cosmetics shop has 4,200 customers with two or more orders. It builds a 1% lookalike from that list — roughly 40,000 people in the country — and spends €300 against it. Result: 45 orders, €6.67 of ad spend per order. The same €300 against a broad interest audience returns 28 orders, at €10.71 each.

The difference is not the algorithm, it is the source list. Had the shop uploaded every registered account, including the ones that never bought, the lookalike would have found people who register and do not buy. A mirror returns whatever you hold up to it.

Why it matters for a business

A lookalike is how a campaign grows past the people who already know you without going back to guessing at interests. It earns its place when remarketing runs out: the visitor list is small, the budget has to increase, and interest targeting has already been tried.

There is a legal side. Uploading a customer list is processing personal data: you need a lawful basis under the GDPR, a privacy notice that allows this use, and a list with the people who opted out removed. The data is hashed before it is sent, but the obligation stays with you rather than with the platform. Google removed similar segments from its ads system in 2023; the equivalent work there is done by optimised targeting on top of your own data.

How to improve it

  • Build from buyers, not from all visitors — and for a small shop, from buyers above your average order value.
  • Give it a few hundred people from one country. Meta’s floor is 100, but results at that size are unstable.
  • Refresh the source. Customers from three years ago do not describe today’s buyer.
  • Test 1% against 3% separately. Broader is not worse once the budget is large.
  • Exclude existing customers so you are not paying twice for the same people.

Source lists and audiences are maintained as part of social media management.

A term you do not recognise? Write to us and we will add it.

The glossary grows with the questions we are asked.