Customer Health Score
A Customer Health Score is a composite metric that predicts a customer's likelihood to renew, expand, or churn — based on product usage, support interactions, engagement, and sentiment data. It's the early warning system that lets CS teams intervene before customers decide to leave.
A SaaS company builds health score: Usage (40% weight) — DAU/MAU ratio normalized to 0-100; Support (25% weight) — inverse of ticket volume + CSAT; Engagement (20% weight) — QBR attendance + feature adoption; NPS (15% weight) — latest NPS score. An account scores: Usage 75, Support 60, Engagement 80, NPS 70 = (75×0.4) + (60×0.25) + (80×0.2) + (70×0.15) = 30 + 15 + 16 + 10.5 = 71.5. Anything under 60 triggers CSM intervention.
Customer health scores are either predictive or performative — there's rarely an in-between. The performative ones: someone picked 10 metrics, weighted them arbitrarily, and now CS ignores the scores because they don't predict anything.
The predictive ones: built from actual churn data. You analyze what churned customers did 3-6 months before cancelling (usage drop, support tickets up, NPS decline, executive sponsor left) and weight accordingly. Then you validate: do accounts with scores under 50 actually churn at 3x the rate?
Here's the RevOps angle: health scores require data from product, support, billing, and engagement systems — unified in one place. Most companies can't build health scores because their data is siloed. Fix the plumbing first, then build the model.
The best health scores aren't complex. Usage trend + support sentiment + engagement frequency usually gets you 80% of the predictive power.
Define ItOther Definitions
“A customer health score is a predictive, data-driven metric that consolidates multiple inputs into a single score to help Customer Success teams evaluate the likelihood of renewal, growth, or churn.”
“Customer health score is a metric used to measure and predict customer satisfaction and engagement. It aggregates usage data, support interactions, NPS, and other signals to indicate whether a customer is at risk or primed for growth.”
“A customer health score is a quantified representation of how healthy or engaged a customer is, combining product usage, relationship quality, and value realization into a single actionable metric.”
A customer health score predicts renewal likelihood by aggregating multiple data inputs into a single metric. Gainsight emphasizes the predictive nature: it evaluates future behavior (renewal, expansion, churn). ChurnZero lists the typical inputs: usage, support, NPS. Totango adds the value realization dimension: is the customer actually getting value?
Common health score components include: (1) Product Usage — feature adoption, login frequency, depth of use; (2) Support — ticket volume, resolution satisfaction, escalation rate; (3) Engagement — executive sponsor activity, QBR participation, training completion; (4) Sentiment — NPS/CSAT scores, survey responses, CSM sentiment; (5) Financial — payment history, contract expansion, service downgrades.
Scoring models range from simple traffic light (red/yellow/green) to weighted 0-100 scales. The key is validation: does the score actually predict outcomes?
MistakesCommon Mistakes
Weighting metrics based on intuition rather than correlation with actual churn
Too many inputs making the score un-actionable (what do you fix first?)
No data infrastructure connecting product/support/engagement data
Setting thresholds arbitrarily instead of validating against outcomes
Using health scores for reporting instead of triggering CS actions
Churn catching you by surprise?
We build health scores from actual churn data — so CS knows which accounts need intervention before it's too late.
Experience across
