Churn Risk Scoring

Churn risk scoring assigns a numerical probability to each account's likelihood of churning within a defined time horizon — typically 90 days and 180 days. The score is calculated from the health model dimensions (from BI-301h), the active churn signals detected, the matched churn archetype (if any), and the environmental risk factors. The score is not a prediction — it is a calibrated probability that enables resource allocation. An account with a 70% churn risk in the next 90 days warrants executive intervention. An account with a 15% churn risk warrants enhanced monitoring. The score determines the response intensity.

1. Calculate Base Risk from Health Score The health model provides the baseline churn risk. A health score below 40 maps to a base churn risk above 50%. A health score above 75 maps to a base churn risk below 10%. The mapping is calibrated against historical data: at each health score level, what percentage of accounts actually churned? The mapping function converts health scores into probability estimates.
2. Adjust for Active Signals Each detected churn signal adjusts the base risk upward. The adjustment magnitude depends on the signal's historical predictive reliability. A Silence Zone entry adjusts risk by +15 to +25 points. A competitor displacement signal adjusts by +10 to +20 points. A leadership transition adjusts by +5 to +15 points. The adjustments are additive up to a ceiling of 95% — no model should claim certainty.
3. Adjust for Environmental Factors Environmental signals that are not captured in the health model provide additional adjustment. Industry-wide budget pressure, macroeconomic conditions, regulatory changes, and competitive landscape shifts all affect churn probability at the portfolio level. These adjustments are applied across cohorts, not individual accounts — a recession increases churn risk for all accounts, not just the ones showing behavioral signals.