Customer Health Score
A composite metric that combines product usage, engagement, support activity, and relationship signals into a single score predicting whether a customer is likely to renew, expand, or churn.
Health Scores Catch Churn Before the Cancellation Email
By the time a customer emails to cancel, the decision was made weeks or months ago. The product usage dropped. The champion left the company. Support tickets went unanswered. A health score surfaces these warning signs early enough to intervene — when the account is trending down, not after it has flatlined.
The best customer success teams use health scores to prioritize their time. In a book of 50-100 accounts, you cannot give equal attention to every customer. Health scores tell you which accounts need proactive outreach, which are thriving, and which need an escalation path.
Building a Health Score Model
| Signal | Weight | Red Flag Threshold |
|---|---|---|
| Product usage (DAU/WAU) | 30% | Below 50% of onboarding month |
| Feature adoption | 20% | Using fewer than 3 core features |
| Support tickets | 15% | 3+ open tickets or negative sentiment |
| NPS/CSAT | 15% | Score below 7 |
| CSM engagement | 10% | No response to last 2 outreach attempts |
| Contract signals | 10% | Payment 30+ days late, no renewal discussion |
Combine into a simple traffic-light system: Green (score 8-10, healthy), Yellow (5-7, needs attention), Red (1-4, at risk of churn).
Making Health Scores Drive Action
A health score without a playbook is just a number. Define specific actions for each tier. Red accounts get an executive sponsor call within 48 hours, a usage review, and an escalation plan. Yellow accounts get a proactive QBR and a success plan refresh. Green accounts get expansion conversations and referral asks. Track save rate — what percentage of red accounts are you pulling back to green? If your save rate is below 30%, either your playbooks need work or you are catching accounts too late. The best teams achieve 40-60% save rates by intervening when the score first dips, not after multiple months of decline.
Frequently Asked Questions
What inputs go into a customer health score?
The most common inputs, weighted by predictive power: product usage frequency (30%), feature breadth (20%), support ticket volume and sentiment (15%), NPS or CSAT score (15%), engagement with CSM and communications (10%), and contract/billing signals like payment delays (10%). The weights vary by product — for a daily-use tool, usage frequency matters more. For a quarterly-use tool, feature depth matters more.
How do you build a customer health score?
Start simple with 3-4 inputs you already track: login frequency, key feature usage, support ticket count, and NPS score. Score each on a 1-5 scale and average them. Test the score against actual churn data — does a low health score predict churn? If it does, refine the weights. If it does not, change the inputs. Most teams iterate through 3-4 versions before the score becomes reliably predictive.
