CX-201a · Module 1

Designing the Health Score Model

4 min read

A health score is a composite signal that tells you where to look. It does not tell you what to do — that is what the conversation tells you. But the score determines which conversations to have, with which clients, and how urgently. A well-designed health score model catches declining accounts weeks before the client sends the cancellation email. A poorly designed one gives you green dashboards while clients quietly churn.

The architecture of a health score model has three components: the pillars (what you measure), the weights (how much each pillar matters), and the thresholds (what score levels trigger what actions). Get the pillars wrong and you are measuring noise. Get the weights wrong and you are prioritizing the wrong signals. Get the thresholds wrong and you are either drowning in false alarms or sleeping through real crises.

  1. Pillar 1: Engagement Frequency and depth of client interaction. Metrics: meeting attendance rate, response time to your communications, number of stakeholders actively participating, client-initiated contact frequency. Engagement is the behavioral signal — it tells you whether the client is investing attention in the relationship.
  2. Pillar 2: Adoption How deeply the client has embedded your deliverables into their workflows. Metrics: active user count versus licensed users, feature utilization breadth, integration depth with existing systems, self-service usage. Adoption is the stickiness signal — low adoption means easy exit.
  3. Pillar 3: Outcomes Whether the client is achieving the results that justified the investment. Metrics: KPI improvement against baseline, time-to-value achievement, stated goal progress. Outcomes are the truth signal — everything else is a proxy for this.
  4. Pillar 4: Relationship The qualitative health of the human connection. Indicators: NPS or CSAT scores, QBR engagement quality, willingness to provide references, tone of communications. Relationship is the sentiment signal — the emotional layer that quantitative metrics cannot capture.