A customer health score is a single metric that measures how much value a customer is getting from your product—and how likely they are to renew, expand, or churn.
For many SaaS companies, it's the difference between reactive support (waiting for support tickets) and proactive retention (knowing who's at risk before they click "cancel").
But here's the challenge: health scores are easy to get wrong. Many companies create scores that don't actually predict churn, change too often to be actionable, or require so much manual work that no one uses them.
In this guide, I'll walk you through a proven framework for building a health score that actually works—including how to calculate it, segment your customers, and turn your score into retention action.
Why Customer Health Scores Matter
Before you build a health score, let's be clear on what you're trying to solve:
- Early warning system: Without a health score, you only know a customer is at risk when they contact support or stop using your product. A health score gives you 30-60 days of warning.
- Prioritization: Your team can't help every customer equally. A health score tells you who needs attention most—and why.
- Automation trigger: Health scores feed into win-back campaigns, cancellation flow interventions, and automated offer delivery (like ChurnZap does).
- Accountability: Customer Success teams can track whether their interventions actually improve health over time.
- Predictive power: A well-built health score correlates with churn, expansion, and NPS.
The ROI is substantial. Companies that implement health scoring see 15-30% faster churn detection and 20-40% better intervention success rates.
The Three Categories of Health Score Metrics
Every effective health score combines three types of data:
1. Product Engagement (Usage Signals)
How much value is the customer actually getting from your product?
- Daily/Weekly/Monthly Active Users (DAU/WAU/MAU)
- Feature adoption rate (% of key features used)
- Time-to-value completion (onboarding steps finished)
- Session frequency and duration
- API call volume (for API-driven products)
- Data processed or results generated
Why it matters: Usage is the strongest predictor of churn. Customers who aren't using your product will churn, no matter what their contract says.
Weight: 40-50% of your overall score (this is your primary signal)
2. Customer Sentiment (Voice of Customer)
What does the customer actually think about your product and relationship?
- Net Promoter Score (NPS)
- Customer Satisfaction (CSAT)
- Recent support ticket sentiment
- In-app feedback/ratings
- Renewal intent (from conversations)
Why it matters: Sentiment catches issues that usage data might miss. A customer might be using your product but hate it. Their sentiment is a leading indicator of churn risk.
Weight: 25-30% of your overall score (secondary signal)
3. Business Metrics (Account Health)
Is the customer healthy from a business perspective?
- Payment status and history
- Support ticket volume and escalations
- Contract renewal date proximity
- Expansion opportunities (unused features, adjacent products)
- Account growth (team size, data usage)
Why it matters: Business metrics contextualize usage. A customer with declining usage + escalating support tickets + upcoming renewal is your highest-risk account.
Weight: 20-25% of your overall score (tertiary signal)
Building Your Health Score: Step-by-Step
Step 1: Choose Your Metrics (5-8 Total)
The best health scores are simple, not comprehensive. Most companies use 5-8 metrics total:
| Metric | Data Source | Weight | Update Frequency |
|---|---|---|---|
| Monthly Active Users (MAU) | Your product database | 25% | Daily |
| Feature Adoption Rate | Your product database | 15% | Daily |
| NPS Score | Survey (quarterly) | 20% | Quarterly |
| Support Escalations (30d) | Help desk | 15% | Daily |
| Days to Renewal | CRM/Billing system | 15% | Daily |
| Recent Support Sentiment | Help desk | 10% | Daily |
Step 2: Normalize Metrics to a 0-10 Scale
Each metric should be normalized so they can be combined. Here's how:
Normalized Score = (Actual Value / Target Value) × 10
Example: MAU Normalization
- Customer has 25 active users
- Your target is 50+ users for "healthy"
- Score = (25/50) × 10 = 5/10
For negative metrics (lower is better):
Normalized Score = 10 - ((Actual Value / Target Value) × 10)
Example: Support Escalation Normalization
- Customer has 3 escalations in 30 days
- Your target is 0 escalations
- Score = 10 - ((3/1) × 10) = capped at 0/10
Step 3: Weight and Combine Metrics
Multiply each normalized metric by its weight, then add them up:
(MAU × 0.25) +
(Feature Adoption × 0.15) +
(NPS × 0.20) +
(Support Escalations × 0.15) +
(Days to Renewal × 0.15) +
(Support Sentiment × 0.10)
= Final Score out of 10
Result: A single number between 0-10 that reflects overall customer health.
Step 4: Define Health Categories and Thresholds
Don't just track scores—segment customers into actionable categories:
| Category | Score Range | Meaning | Action |
|---|---|---|---|
| Healthy | 8.0 - 10.0 | Customer is thriving, getting value, engaged | Nurture for expansion; gather testimonials |
| At Risk | 5.0 - 7.9 | Customer showing warning signs; declining engagement or sentiment | Proactive outreach; offer support or resources |
| Critical | 0.0 - 4.9 | Customer at high churn risk; major issues | Executive outreach; retention offer; intervention playbooks |
Step 5: Segment by Business Model and Customer Size
One health score doesn't fit all. Create different scoring models for different segments:
SMB Score (Self-serve users):
- MAU (40%) - usage is primary driver
- Feature Adoption (30%)
- In-app NPS (20%)
- Support tickets (10%)
Enterprise Score (CSM-managed):
- Onboarding Completion (15%) - slow ramp, so weight differently
- Executive Engagement (25%) - who's using it at the top
- Feature Adoption (20%)
- Expansion Revenue (20%) - are they using more?
- CSM Assessment (20%) - human judgment on strategic fit
Real-World Implementation: The Acme SaaS Example
Let's walk through how a real company implements health scoring:
Current State (Before Health Scoring):
- 30% churn rate (unacceptable)
- Customer Success team finds out customers are leaving when they get cancellation requests
- No way to prioritize which accounts need help
- Retention efforts are random and reactive
Health Score Implementation:
- Week 1-2: Audit their data. They have product usage data, NPS surveys (done annually), and support ticket volume.
- Week 2-3: Build the health score with 5 metrics: MAU (30%), Feature Adoption (20%), NPS (25%), Support Tickets (15%), and Contract Renewal Proximity (10%)
- Week 3-4: Sync health scores with their CRM. Create three customer segments: Healthy (100 accounts), At Risk (45 accounts), Critical (8 accounts)
- Week 4+: Set up automated alerts when a customer moves into "At Risk" or "Critical"
Results (After 90 Days):
- CSM team reaches out to all 45 "At Risk" customers proactively
- Discover that most are stuck on onboarding (issue they didn't know existed)
- Implement onboarding improvements; 28 of 45 accounts improve to "Healthy"
- 8 "Critical" accounts get special attention; prevent 5 from churning through targeted retention offers
- Net result: Churn drops from 30% to 22% in 90 days
Common Health Score Mistakes to Avoid
- Making it too complex: Health scores with 15+ metrics become expensive to maintain and confusing to act on. Stick with 5-8.
- Ignoring recent trends: A score is a snapshot. What matters is whether a customer's score is improving or declining. Set up trend alerts.
- Using stale data: An NPS survey from last quarter won't help you today. Pair infrequent metrics (like NPS) with real-time signals (like usage).
- Not involving your team: Build health scores with input from CSM, Product, and Support teams. They have insights your data doesn't.
- Treating all metrics equally: Usage matters more than support ticket count. Weight metrics based on their correlation with actual churn in your business.
- Setting unrealistic thresholds: Your "Healthy" threshold should reflect actual customer behavior, not wishful thinking.
Tools for Health Score Calculation
You have options depending on your team's technical depth:
| Tool Category | Best For | Examples |
|---|---|---|
| Native Platform Features | Quick setup, all-in-one solution | Gainsight, Totango, Vitally, Planhat |
| Analytics Platforms | Deep product usage insights | Amplitude, Mixpanel, Heap |
| Data Warehousing | Custom scoring, full control | Segment, Rudderstack, custom Python/SQL |
| Spreadsheet-based | Proof of concept, small datasets | Google Sheets, Excel with formulas |
Recommendation: Start with native platform features (Gainsight or Vitally if you have budget; Segment + spreadsheet if you don't). Once you've proven the concept and understand your weighting, consider moving to a full data warehouse solution.
Connecting Health Scores to Retention Action
Here's where most companies fail: they build a health score but don't actually use it.
A health score is only valuable if it triggers action. That means:
- Automated alerts: When a customer moves to "At Risk," CSM gets alerted immediately
- Automated interventions: "Critical" accounts automatically get offered a retention discount, win-back email, or other retention tactic (this is where ChurnZap excels)
- Dashboard visibility: Your entire team should see health scores at a glance
- Playbook assignment: Different health states trigger different playbooks (onboarding support for new customers, advanced training for power users, etc.)
Key Takeaway
A customer health score transforms churn from inevitable to preventable. Instead of reacting to cancellations, you're proactively identifying customers at risk—and getting your team in front of them before they decide to leave.
Start simple (5-8 metrics), validate that your score correlates with churn, and improve from there. Most companies see measurable churn reduction within 60 days of implementing a health score.
Ready to Automate Your Retention Response?
A health score identifies at-risk customers. ChurnZap automatically prevents them from churning by deploying intelligent retention offers in the cancellation flow.
Join the Waitlist