Health Score Builder

A weighted health score framework, tested against representative accounts.

This framework is independent of the churn risk calculator's: inverted scale (here 0 to 100, a high rating is a good signal; there 0 to 4, a high rating is a deterioration), its own families and weights, and thresholds that do not line up. No score converts into the other, and the two are deliberately not harmonised.

Family weights

Reading convention: a high rating is a good signal, support included, where 90 means few tickets and well handled.

Band thresholds

Vetoes

A veto bars the green band whatever the score: amber at best. An account whose sponsor has left, or with an open dispute, should not read as green.

Representative account profiles

Representative account profiles, one profile per row, a rating from 0 to 100 per family
Profile Usage Support Relationship Contractual Remove

Ratings from 0 to 100, one per family. Entries stay in the browser and disappear when the tab is closed. Use a point for decimals, e.g. 1.5.

Average profile score
0.0
Distribution by band
0
reading

Profiles run through the framework

Score, band and deciding signal of each profile
Profile Score Band Deciding signal

Average contribution by family

Framework summary

.
What this calculation does not say
1. The framework remains a causal hypothesis, which no renewal history has confirmed or contradicted.
2. Logins, sessions and active features are heavily correlated: adding them counts one piece of information three times and drowns out the relationship and contractual families.
3. The thresholds are calibrated on the current shape of the book: a wave of small accounts shifts them without a single line of the framework having moved.
4. The score rates a state, never a direction: an account sitting stably at the bottom of a band and one falling for three months show the same value.
5. This framework cannot be compared with the churn risk calculator's: inverted scales, different weights, thresholds that do not line up. A green account here is not necessarily green there.