Product & Analytics / Data Science
Churn-Prediction / Propensity Score ~ (also "propensity score", "churn risk score")
A model's estimate of how likely a given user is to churn (or convert, or resurrect) in an upcoming window.
The fork, why two teams get different numbers
WHAT the score means and HOW it is cut.
In the Metric Library
- The full fork, both definitions worked all the way through
- The trap that makes the number lie, on a real export
- Every formula variant, spelled out
- The reconciliation anchor, what to tie it to and when to refuse
Reference: Applied-ML convention (no governing body); precision/recall and probability-calibration (reliability-curve) methods cited as reference