A health score is a combined indicator that helps a SaaS company estimate the likelihood that a customer is healthy, at risk, or in need of support. It can bring together signals such as product usage, feature adoption, support activity, customer feedback, payment status, and engagement with the customer success team.
A health score is not a universal measure of customer satisfaction or a guaranteed prediction of churn. It is most useful as an early-warning and prioritization tool, supported by conversations with customers and context from the teams that serve them.
How Does a SaaS Health Score Work?
Teams select signals that are meaningfully connected to customer outcomes, decide how each signal should affect the score, and combine them into a rating or category. Some companies use a numeric score, while others use labels such as healthy, watch, and at risk.
For example, a team might consider whether a customer logs in regularly, uses key features, completes onboarding, has unresolved support issues, and reports progress toward their goals. The importance of each signal depends on the product and customer segment.
Example Health Score Model
Product usage: 40 points
Adoption of key features: 25 points
Customer feedback: 20 points
Support and account signals: 15 points
Total possible score: 100 points
This is an example structure, not a standard formula. A company should test whether the selected signals actually help identify customers who need attention. Some signals may need to lower a score, raise it, or trigger a separate review rather than being added as simple points.
Why Health Scores Matter
Prioritize customer success work
When a team supports many accounts, a health score can help identify which customers may need proactive outreach. It can help focus attention, but it should not replace human judgment or direct customer feedback.
Spot risk earlier
A drop in usage, lack of progress during onboarding, unresolved issues, or negative feedback may indicate that a customer is struggling. Acting early gives the team a chance to understand the problem and offer relevant help.
Guide adoption and expansion
Health signals can also highlight customers who are achieving value and may benefit from additional training, workflows, or product capabilities. Expansion conversations should still be based on customer needs and fit.
Support consistent account reviews
A shared model can help customer success, support, product, and sales teams discuss account status using a common starting point. Clear definitions reduce confusion and make follow-up actions easier to assign.
Signals Commonly Used in a Health Score
| Signal | What it may indicate | Potential limitation |
|---|---|---|
| Product usage | Whether users are returning and using the service. | Login frequency alone may not reflect meaningful value. |
| Key feature adoption | Whether customers use features tied to their goals. | Different customers may need different features. |
| Onboarding progress | Whether setup and early milestones are completed. | Complex deployments may take longer for valid reasons. |
| Support activity | Possible friction, recurring issues, or unresolved blockers. | Many tickets can reflect strong engagement or a complex use case, not only risk. |
| Customer feedback | Reported satisfaction, confidence, or concerns. | Feedback may be missing or come from only a few contacts. |
| Commercial signals | Renewal status, payment issues, or changes in contract use. | Billing or procurement events need context and careful handling. |
How to Build a Useful Health Score
- Define the outcome. Decide whether the score should help predict churn, identify onboarding blockers, guide customer success outreach, or support another clear action.
- Start with a small set of signals. Prefer indicators with a clear relationship to customer outcomes over data that is merely easy to collect.
- Segment customers. Enterprise accounts, small businesses, and self-serve customers may show health differently. Avoid assuming one model fits everyone.
- Set transparent rules. Document how signals are scored, weighted, updated, and interpreted.
- Validate against outcomes. Compare historical scores with renewals, churn, adoption, and customer feedback. Check for false alarms and missed risks.
- Combine quantitative and qualitative evidence. Account notes and direct conversations can explain why a score changed.
- Assign follow-up actions. A score should point to a useful next step, such as resolving a blocker, offering training, or contacting the customer.
- Review and improve the model. Update it when the product, customer mix, or evidence changes. Monitor whether the score remains useful.
Health Score vs. Customer Health Score vs. NPS
| Metric | Main purpose |
|---|---|
| Health score | A combined signal used to assess a customer’s current status or likely needs. |
| Customer health score | A health score specifically designed to assess the relationship or success of an individual customer or account. |
| Net Promoter Score (NPS) | A survey-based measure of how likely respondents are to recommend a product or company. |
| Customer satisfaction (CSAT) | Feedback about satisfaction with a particular experience, interaction, or product. |
Common Health Score Mistakes
- Using arbitrary weights without checking whether the score predicts useful outcomes.
- Treating low usage as a problem for every customer, even when their workflow requires infrequent use.
- Counting support tickets as negative without considering severity, resolution, or customer context.
- Failing to update stale or missing data.
- Using one score for every segment despite different customer journeys.
- Assuming a green score means no outreach is needed or a red score means churn is certain.
- Collecting or using customer data without appropriate transparency, access controls, and privacy safeguards.
Frequently Asked Questions
What is a good customer health score?
There is no universal score or threshold. The right model depends on the product, customer segment, and outcome the team wants to predict. Define categories based on evidence and validate them against actual customer outcomes.
How often should a health score update?
Update it as often as the underlying signals and customer decisions require. Product usage may update frequently, while survey feedback or account reviews arrive less often. Show when the score was last updated.
Can a health score predict churn?
It may help identify patterns associated with churn, but it cannot predict every outcome. Validate the model against historical data and use direct customer conversations to understand risk.
Who should own the health score?
Customer success or revenue operations may coordinate the model, while product, support, sales, and data teams help choose signals and validate results. The people acting on the score should understand its limits.
Should customers see their health score?
That depends on the purpose and design of the model. Regardless of visibility, teams should be transparent about relevant data practices and avoid treating an internal score as an unquestionable judgment about a customer.