Ad-tech
Lookalike audience
Algorithmic audience built to resemble a seed of high-value users.
A lookalike audience is built by an ad platform's ML model to resemble a seed list of high-value users (purchasers, repeat buyers, free-trial-converters). Seed quality matters more than seed size — a 500-purchaser seed often outperforms a 10K-website-visitor seed. Tighter percentages (1% LAL) are higher-quality but smaller; broader (5–10%) sacrifice fit for reach. Meta's lookalike is the most-tested; Google's Customer Match performs similarly with a different LAL implementation.
Example
DTC supplement brand uploads its top-quartile-LTV customer list (8,400 emails) as a Meta seed; 1% LAL produces 2.1M reach in the US; CPA on the LAL audience is 28% lower than broad-targeting baseline.