Any type of activity such as on-site browsing and conversions are indicators of engagement and therefore play a role in how we score users from Active to At Risk to Churned.
Definition of Churn Score: A statistical machine learning prediction value that estimates the state of churn for a given user at any given time, evaluating user demographic info, browsing behavior, and historical purchase data among other signals, and factors in our unique and proprietary predictions for how long a user will be a customer. Updated daily for all users who have a minimum of one conversion. These values are supplied in three bands: high, medium, and low Churn Score for campaign targeting and reporting. Values can also be aggregated to the entire user base to arrive at a network-level Churn Score.
Definition of Churn Time: A statistical machine learning prediction estimate of the number of days from the current date that a given user is likely to cease making any future purchases. This model evaluates time data from user browsing and purchase behavior (e.g. sign up time, last time made a purchase, last time browsed site/app). The values are aggregated into three cohorts: high, medium, and low Churn Time for campaign targeting and reporting. Values can also be aggregated to the entire user base to arrive at a network-level Churn Time.