Lookalike audience

What it is

You supply a seed: a customer list, a website or app event audience, video viewers, or people who engaged with your page. The platform models what those people have in common, ranks the wider user pool by similarity and offers the top slice as a target.

On Meta the slice is a percentage of a country's population: 1% is the narrowest and most similar, 10% the broadest. Google's closest equivalents are Customer Match expansion and similar segments; LinkedIn, TikTok and Microsoft Advertising ship their own versions of the same idea.

Why it matters

Seed quality decides the outcome. A few thousand genuinely valuable customers produce a better model than a hundred thousand undifferentiated visitors; the model learns whatever pattern you give it. Build the seed from your best customers and keep it current.

Size is a trade-off: narrow percentages bring higher precision at higher CPMs, broad ones bring cheap reach but drift toward generic targeting. Common practice is to run 1-2% and 5-6% variants as separate ad sets and let performance decide. Every uploaded list needs explicit consent and a lawful basis under GDPR.

How Tecrube uses it

On Meta Ads you create website and lookalike audiences, search interests and estimate reach from the chat; on Google Ads you manage audience signals and Customer Match lists.

Creating an audience is a write action. Tecrube previews which seed, which region and which size the audience will use; nothing is sent to the platform without your approval.

Example

“Build 2% and 5% lookalikes for the United Kingdom from customers who bought twice in the last 180 days, ready for separate ad sets.” The assistant previews both audience definitions; on approval they are created on Meta and logged.