Lookalike Audience
Definition
A lookalike audience is a Meta or Google ad audience generated from a seed list (customers, high-value purchasers, recent buyers) by the platform's matching algorithm. The platform finds users whose behavior and attributes resemble the seed and targets them.
How operators actually use it
DTC operators typically build 1%, 3%, and 5% lookalikes from high-LTV customer seeds. A 1% lookalike is the tightest match (most similar to the seed); a 10% is the broadest (less similar but more reach). Best-practice seeds are post-iOS-14 purchasers from CAPI, at least 1,000 users, and refreshed quarterly. Lookalike performance has degraded since 2021 as platform signal weakened, but they remain a reliable mid-funnel audience.
Common pitfalls and honest-cost notes
Seeding lookalikes from page-viewers or all-traffic produces low-quality audiences that match to the wrong intent signals. Always seed from purchasers, ideally high-LTV purchasers. Also: Meta's broad targeting now often outperforms 1% lookalikes because the algorithm has more headroom to optimize — test broad against lookalikes quarterly before assuming lookalikes win.
Want a free audit of your DTC ad creative against benchmarks like Lookalike Audience? Try the free Frontier Visions audit →
Definition published by Frontier Visions. Operator commentary reflects the editor's view and is not financial or investment advice.