Lookalike Modeling is comparing a seed audience (either a Retargeting or Conversion audience pool) with the universe of 3rd party audience pools provided by 3rd party audience data vendors and searching for different audiences most statistically alike to the seed. It identifies data segments most relevant, with the highest correlation of similarity, and best suited for prospecting.
The system looks at what 3rd Party Data Elements (across all data providers) are associated with the unique users on the pixel and determines the Relevancy Score for each Data Element.
How large of a seed audience is necessary?
The minimum seed audience size for lookalike modeling is 30,000 uniques over 30 days to ensure the model is statistically significant.
Where does the 3rd party audience data come from?
Audience is sourced from integrations with over 30 different audience data partners, including the below.