strategist-predictions

Configure predictive churn scoring by defining your churn prediction windows, selecting your target audiences, and setting score thresholds to segment at-risk users. This calibrates your predictive model to accurately identify users likely to churn before they leave.

Churn Risk Score predicts the likelihood of a user leaving, while Event Likelihood Score predicts the probability of a specific conversion or purchase event. Both scores categorize users into tiers for targeted retention or conversion campaigns and Canvas flows.

Build a retention Canvas by segmenting users based on a Churn Risk Score threshold, such as a score above 70, and enrolling that high-risk cohort into a tiered re-engagement flow to deliver timely messaging strategies.

Predictive model training fails or shows poor prediction quality due to data-quality errors or insufficient training data within your selected prediction audiences. Troubleshoot by validating your audience constraints and ensuring enough historical data exists for the prediction window.

Yes, you can use predictive event likelihood for targeted upsell campaigns by segmenting users with high Likelihood Category scores. This allows you to trigger conversion campaigns and purchase prompts specifically for users with high intent.

Predictive modeling training constraints involve audience size limits and specific prediction window durations that must be met to generate a valid model. Understanding these constraints ensures accurate analytics interpretation and prevents common training errors.