Look-alike modeling

A technique for identifying groups of people who share characteristics and behaviours similar to a brand's best-performing customers.

The process starts with a defined seed audience — typically a brand's most profitable or engaged customers — and uses algorithmic modeling to find new users who resemble that group across relevant signals such as demographics, browsing behaviour, and purchase patterns. The goal is audience expansion without sacrificing relevance.

For example, if an online retailer's highest-value customers are those with an average purchase over $100, an interest in cosmetics and perfumes, and a purchase frequency of at least twice a month, look-alike modeling can surface a broader pool of prospective buyers who match that profile — enabling the advertiser to reach net-new users with a strong likelihood of converting.