Lookalike audience
An ad audience a platform builds by finding people who resemble a source list you supply.
A lookalike audience is generated by an ad platform from a source audience you provide: customers, site visitors, people who engaged with a video. The platform models what that group has in common and finds more people like them. Meta documents its own implementation in the Meta Business Help Center.
The output is only ever as good as the input. A source list of your best customers produces a genuinely useful audience; a source list of everyone who ever landed on your homepage produces something close to broad targeting with extra steps.
Treat the mechanism as a black box you can't audit. You supply a list, you get an audience, and you cannot see the reasoning. That's a reason to test it against a plain broad-targeted set rather than to assume it wins.
Where this comes up
The organic post proved the creative. It didn't prove the targeting, the offer or the destination. What transfers, what doesn't, and how to structure the build.
Why the old even-thirds funnel split stopped making sense, what a modern default looks like, and how to know when yours is wrong.