Back to Blog

Building a Customer Health Score Dashboard

Design a customer health score that predicts churn and identifies at-risk accounts. Use it to save customers before they leave.

March 24, 2026Dashboard & ReportingMetricGen Team

Customer success teams are firefighters. They react to problems.

A customer health score makes them paramedics. It predicts problems before they happen.

This guide covers how to build and use a health score dashboard.

What Is a Customer Health Score?

A single number (0-100) predicting: "Will this customer renew?"

High score (80+) = likely to renew Low score (40-) = at-risk, might churn

How to Calculate It

Weight the factors that predict churn:

Product Engagement (30%)

  • Login frequency
  • Feature adoption
  • Sessions per week
  • Days since last login

Usage Growth (25%)

  • Seats/users growing?
  • Usage volume trend
  • Feature expansion

Support Health (20%)

  • Open support tickets?
  • Response times met?
  • Sentiment from tickets

Business Health (25%)

  • Payment issues?
  • Contract renewal date
  • Payment rate (paying on time?)

Formula:

Health Score =
  (Engagement × 0.30) +
  (Growth × 0.25) +
  (Support × 0.20) +
  (Business × 0.25)

Scale 0-100

The Dashboard Layout

┌─────────────────────────────────────────┐
│ AT-RISK CUSTOMERS (score < 40)         │
│ Customer A: 25 (no login in 30 days)   │
│ Customer B: 38 (2 open support tickets)│
│ [ACTION: Reach out today]              │
└─────────────────────────────────────────┘

┌──────────────────────────────────────────┐
│ NEEDS ATTENTION (score 40-60)           │
│ Customer C: 55 (engagement down 40%)    │
│ Customer D: 48 (ticket response slow)   │
│ [ACTION: Check in this week]            │
└──────────────────────────────────────────┘

┌──────────────────────────────────────────┐
│ HEALTHY (score 60+)                     │
│ Customer E: 88 (usage up, engaged)      │
│ Customer F: 92 (expanding seats)        │
└──────────────────────────────────────────┘

[Trend: Health scores down 5% from last month]

Color-Coded Risk Levels

  • 🔴 Red (0-40): Critical, reach out immediately
  • 🟡 Yellow (40-60): At-risk, schedule check-in
  • 🟢 Green (60-100): Healthy, nurture for expansion

Key Features

1. Trending: Show score over time. A dropping trend is a warning even if current score is decent.

2. Drill-Down: Click a customer to see: "Why is their score low?"

  • Low engagement? (Missing feature tutorials?)
  • Support issues? (Fix those tickets)
  • Payment delays? (Contracts upcoming?)

3. Segmentation: Filter by:

  • Industry / segment
  • Customer size
  • Product tier
  • Tenure

This reveals patterns. Maybe Enterprise customers always dip in Q2.

Using the Dashboard

Daily (CS team):

  • Check red/yellow customers
  • Reach out to at-risk accounts
  • Log actions taken

Weekly (Manager):

  • Trend analysis (is health improving?)
  • Identify patterns (which segments are at-risk?)
  • Allocate resources (where to focus CS effort?)

Monthly (Leadership):

  • Correlation analysis (is product improving retention?)
  • Forecast (predict churn for next quarter)
  • Strategy (which customers need improvement? Which markets?)

What a Good Health Score Does

✓ Predicts churn 4-8 weeks in advance (time to intervene) ✓ Identifies root cause (engagement? support? pricing?) ✓ Guides CS actions (who to reach out to, what to fix) ✓ Validates product improvements (did feature launch improve scores?) ✓ Segments customers (which deserve enterprise support?)

Building the Score: Frameworks

Simple (DIY): 1-5 scale for engagement, support, growth, business Average them. Done.

Moderate (tools-based): Use a CS platform (Gainsight, Planhat) that tracks these automatically.

Advanced (data-driven): Use ML to weight factors by their actual correlation to churn.

Start simple. You can sophisticate later.

Common Mistakes

Using vanity metrics: Email opens don't predict churn. Actual usage does.

Ignoring external factors: Market downturn affects everyone. Score needs context.

Setting it and forgetting it: Health scores decay. Recalibrate quarterly.

Not acting on low scores: A dashboard is useless if CS doesn't follow up.

Checklist

  • ✓ Health score predicts churn (validate against actual churn)
  • ✓ Factors weighted by importance
  • ✓ Includes engagement, growth, support, business health
  • ✓ Scored 0-100 (easy to interpret)
  • ✓ Color-coded (red/yellow/green)
  • ✓ Shows trend (not just current score)
  • ✓ Drill-down to root causes
  • ✓ Updated at least monthly
  • ✓ CS team acts on low scores
  • ✓ Calibrated against actual churn

The Bottom Line

A health score is a prediction tool. Use it to call customers before they leave.

A customer with a dropping health score is more valuable than data about a customer who already churned.

Get ahead of churn. That's what the dashboard is for.


Explore the full metric definition

MetricGen has chart templates, formulas, and sample data for hundreds of business metrics.

Browse Metrics

Related Guides