You choose a source — a customer list, website purchasers, high-value leads — and a location, and Meta builds an audience of people who resemble it. Sizes range from 1% (the closest match) to 10% of a country's population. Value-based lookalikes, built from lists that include customer value, weight the match toward your most valuable customers.
Lookalike quality depends heavily on the source: a list of your best customers produces far better lookalikes than a list of every website visitor. As Meta's broad targeting has improved, lookalikes are now often used as a suggestion inside Advantage+ audience rather than as a strict boundary.
Example
An HVAC company builds a 1% lookalike from customers who bought a full system replacement, reaching people with similar profiles across its state.
How Sloss Digital uses it
We seed lookalikes from high-value customer segments — not just any converter — and test them against broad targeting to see which finds better customers.