Lookalike audience percentage
Also known as: Lookalike percentage
In short
A size setting for a modeled audience that trades a narrower similarity match against a broader potential audience.
A lookalike audience percentage describes the size of a modeled audience relative to the population used by the platform. The model starts from a source audience and finds other people with similar characteristics. The percentage is not the probability that someone will buy.
What a smaller percentage means
Smaller lookalike audiences concentrate on closer modeled similarities. Larger audiences expand the available pool. Meta’s public examples distinguish these similarity levels and show that percentage-based tiers can be created separately.
A cumulative audience and a separate tier are different selections. A wider cumulative audience can include people from the narrower selection. A tier such as 1-2% is intended to describe the next band, rather than the complete first 2%.
The quality of the source audience remains important: a model based on customers and one based on casual visitors answer different questions.
Illustrative example
A retailer compares a narrow customer-based lookalike with a broader band using the same offer. It checks whether the groups overlap before interpreting the results as a clean comparison.
This is an illustrative test design, not a recommendation to copy a particular percentage. A smaller audience does not guarantee lower acquisition cost. Confirm the controls and delivery behavior available in the actual campaign before treating the chosen percentage as a strict reach boundary.