Analytics
Predictive Audience
A predictive audience is a segment built using machine learning models that identify users most likely to perform a specific action such as purchasing, churning, or upgrading. It moves beyond retrospective behavioral targeting to forward-looking probability-based targeting.
Examples
An ad platform builds a predictive audience of users with a 70%+ probability of purchasing within the next 7 days, allowing the advertiser to focus budget on highest-intent prospects.
Best Practices
Validate predictive audiences against actual outcomes, compare performance against standard behavioral audiences, refresh models regularly, and use prediction confidence thresholds to balance reach and precision.