Lookalike Audience

A lookalike audience is a group of brand new people who resemble a group you already have, built by handing Meta a source and a country and letting it go find more people in that country who look like your source.

The way it works in practice is that you start from a custom audience you already own, say your list of actual purchasers or the people who bought on your website, and you point Meta at it inside Ads Manager. Meta studies what those people have in common, then goes looking across that same country for new people who share those traits. You pick how wide to cast the net using a size slider that runs from 1 percent of the country's Meta users up to 10 percent, where 1 percent is the smaller, most similar group and 10 percent is a bigger, looser one. So a 1 percent lookalike of your purchasers is simply Meta's best guess at the people who most closely resemble the customers you already have.

This matters because it is one of the cleanest ways to reach cold people who have a real chance of buying, since the whole thing starts from people who already did. You are not describing who you think your customer is and hoping you got it right; you are letting Meta reverse engineer your real buyers and find more of them. The more specific and higher quality your source, buyers rather than just anyone who visited, the better the lookalike it builds.

A lookalike is only ever as good as the seed you feed it. Build one from a tiny or weak source, like fifty random newsletter signups, and Meta has too little to learn from, so what comes back is thin. Build it from your best customers and it is strong. Worth knowing too: as Meta's broad targeting and its own systems have got better at finding buyers on their own, lookalikes carry a bit less weight than they once did, but a lookalike of real purchasers is still a solid, high-intent place to start.

Feed it your buyers, not your bystanders, and the lookalike does its job.