Health Scoring

Last updated: 2026-05-18

SaaSy scores every customer on a 0–100 scale by combining five weighted dimensions. The score updates whenever new data arrives from connected integrations.

The five dimensions

Dimension	Weight	What it measures
Usage	25%	Login frequency, session duration, API call volume, data volume trend
Adoption	25%	Feature adoption rate, user activation rate, advanced features used, active integrations
Support	20%	Open ticket count, average resolution time, escalations in last 30 days, satisfaction score
Financial	20%	Payment history, invoice aging, expansion revenue over 12 months, contract compliance
Relationship	15%	Executive engagement, QBR completion rate, NPS score, champion identified

The overall score is the weighted sum of each dimension's score (0–100 per dimension).

Score thresholds and status

Score range	Status	Grade
90–100	Healthy	A+
80–89	Healthy	A
75–79	Healthy	B
60–74	Stable	B or C
50–59	Stable	D
0–49	At Risk	F

The status (Healthy / Stable / At Risk) is what appears on the customer list. The grade (A+ through F) appears inside the customer detail view.

The customer list colour codes are:

• Green — score ≥ 75 (Healthy)
• Yellow — score 50–74 (Stable)
• Red — score < 50 (At Risk)

How each dimension is scored

Usage (25%)

A high usage score requires consistent logins and active API/feature use. Key signals:

• Login frequency — scored 0–100 based on how often users log in relative to their seat count. Login frequency below 30% of expected triggers a "Warning" indicator.
• Session duration — longer average sessions indicate deeper engagement.
• API usage rate — relevant for customers using SaaSy via API (Scale plan).
• Data volume trend — growing data volumes suggest active use; shrinking volumes are a warning.

Adoption (25%)

Tracks whether customers are actually using the product, not just logging in.

• Feature adoption rate — percentage of available features the customer has used at least once. Rate below 40% contributes to churn risk.
• User activation rate — percentage of provisioned seats that have logged in at least once.
• Advanced features used — count of non-onboarding features engaged.
• Active integrations — number of connected data sources. More integrations = higher stickiness.

Support (20%)

High support load combined with low satisfaction is one of the strongest churn predictors.

• Open tickets — more than 5 open tickets triggers a Critical indicator.
• Avg resolution time — relative to your team's SLA baseline.
• Escalations in last 30 days — more than 2 escalations contributes 0.2 to churn risk.
• Satisfaction score — post-ticket CSAT, if collected.

Financial (20%)

Payment reliability and revenue growth signals.

• Payment history — Clean (good) vs. issues (late/failed). Any payment failure adds risk.
• Invoice aging — invoices more than 30 days overdue trigger a Warning indicator.
• Expansion revenue — positive expansion over the trailing 12 months is a protective factor.
• Contract compliance — whether the customer is within their contracted usage limits.

Relationship (15%)

Human engagement signals that are harder to automate but highly predictive.

• Executive engagement — whether senior stakeholders are actively involved (active / passive / none).
• QBR completion rate — percentage of scheduled Quarterly Business Reviews completed.
• NPS score — Promoter (9–10) is protective. Detractor (0–6) adds significant churn risk.
• Champion identified — whether there is a named internal champion.

What triggers a score change

Scores are recalculated when any of the following occur:

• A connected integration syncs new data (most integrations sync every hour).
• You manually trigger a score refresh from the customer detail view.
• A webhook event arrives that changes engagement data (e.g., a Stripe invoice.payment_failed).
• The nightly background job runs (scores at least once per 24 hours even if no events arrived).

Score history is tracked, so you can see the trend over time on the customer detail page.

Interpreting the score in practice

A score of 85 with declining login frequency is more concerning than a score of 70 with improving adoption — because the trend matters as much as the number. Always look at the dimension breakdown, not just the aggregate, when deciding whether to reach out.

The churn prediction model uses the same underlying signals but weights them differently based on behavioral trajectories. See Churn Prediction.