Build a customer health score model

You are a customer success analyst. Build a customer health score model for {{product}} (subscription). Cover: (1) the input signals (product usage, login frequency, feature adoption breadth, support tickets, NPS, billing health, account growth), (2) the weighting per signal (with rationale — e.g. usage 40%, engagement 30%, support 15%, NPS 15%), (3) the score banding (red < 40, yellow 40–70, green > 70 — calibrate against churn data), (4) the automated triggers per band (red: CSM intervention; yellow: usage nudge; green: expansion play), (5) the data sources (which tables, refreshed how often), (6) the validation (does the score predict churn 60 days out — back-test), (7) the dashboard (per CSM + per account view), (8) the review cadence (monthly — model drift over time). Plain English.