Customer health score: what it is and how to calculate it

A customer health score is a number from 0 to 100 that summarizes how engaged a customer is. It helps you spot churn risk and expansion opportunity before renewal week.

Most teams find out a customer is leaving after the decision is already made. A health score gives you a single read on usage trends so you can act while there is still time.

Why health scores matter

Without health scores, customer success is reactive. You hear about problems when customers complain and learn about churn when they cancel.

Teams that monitor scores catch fading accounts earlier. If you can see which customers are quietly disengaging, you can reach out before they are gone.

Health scores also reduce loudest-customer bias. Without data, CS teams often spend time on whoever emails the most, not the accounts that need help. A score surfaces the quiet customer who stopped logging in three weeks ago and never filed a ticket. Those accounts are prime silent churn candidates.

The four signals behind the score

Before any formula, know what you are measuring:

• Activity: Are they showing up? (logins, sessions, event volume)
• Engagement: Are they using it deeply? (features used, session length, workflow breadth)
• Milestones: Are they progressing? (onboarding done, core features activated, team invites)
• Recency: When did they last do something meaningful? (gaps between sessions)

Each signal is scored from 0 to 100, then combined into one number.

How the score is calculated

A strong health score blends all four signals with weights that reflect how well each one predicts churn in your product.

FirstDistro calls this weighted formula the Signal Stack. Activity gets the highest weight because when customers stop showing up, other signals usually follow.

Try the free health score calculator to see how the weights combine.

Understanding each signal

Activity (35% weight) measures how often a customer uses your product: logins, sessions, and daily active usage. When activity declines, cancellation often follows within 30-60 days.

Engagement (20% weight) measures depth: features used per session, workflow completion, and whether usage is broad or stuck on one screen. High engagement means value across the product, not just a single habit.

Milestones (15% weight) track journey progress: onboarding complete, core features activated, teammates invited. Customers who hit milestones tend to stick longer.

Recency (30% weight) measures time since the last meaningful action. Seven days of silence often triggers a noticeable drop because quiet accounts are frequently the first sign of fading usage.

Score thresholds and what they mean

Each score band suggests a different response. Do not wait for "critical" if the trend is already falling.

The critical threshold is 40. Below that, recovery gets harder. Strong CS teams often intervene in the monitor band (40-69), when decay is still early.

How health scores connect to fading usage

Churn is rarely sudden. Usage usually fades in stages over weeks: longer gaps between sessions, shorter visits, narrower feature use, then near silence.

The pattern runs from thriving through coasting, fading, ghosting, and gone. Health scores help you spot which stage an account is in. For the full stage breakdown and intervention windows, read how customers quietly disengage before they cancel.

Recency often drops first, then activity, then engagement, then milestones stall. That sequence is why a single metric like login count can look fine while the score is already falling.

What to do when a score drops

A score without a playbook is just a number. Define who gets notified, what outreach runs, and when to escalate for each band.

A practical loop: detect the trend, score it, alert the owner, intervene with a specific action, then check whether usage recovers. Repeat and refine what works.

For concrete save tactics once an account is at risk, see how to reduce SaaS churn.

Building a health score step by step

1. Choose your signals. Start with activity, engagement, milestones, and recency.
2. Collect data. Track product events; add CRM context if renewal timing matters.
3. Normalize to 0-100. Decide what "perfect" looks like per signal for your product.
4. Apply weights. Use 40/30/20/10 as a starting point; tune against your churn history.
5. Set thresholds and a playbook. Map each band to an owner and action.
6. Monitor and calibrate. Compare scores to actual renewals quarterly.

For normalization methods and CRM extensions, see the health score formula guide.

Common mistakes

• Using only one signal. Login count alone misses shallow sessions. Engagement alone misses infrequent users. Combine signals.
• Equal weighting. Activity usually predicts churn better than milestones. Weight by predictive power, not convenience.
• Ignoring recency. Strong history plus two weeks of silence is not a healthy account.
• Static thresholds only. A score of 72 and dropping can be riskier than 78 and stable. Watch trends.
• Manual spreadsheets. Stale scores miss the window to intervene. Automate updates.
• No playbook. Define what happens at each band before you need it.

Health score vs dashboard

A dashboard lists metrics. A health score compresses behavior into one number and a trend you can sort on Monday morning.

The point is not another green chart. It is knowing which accounts need a call this week. After you intervene, prove the dollars with saved revenue. For why a greener score is not proof, see health scores vs saved revenue. Use net revenue retention for the portfolio scoreboard.

Software comparison

Platforms differ mainly on setup burden: rule-based thresholds you configure by hand versus systems that learn weights from your usage data.

Rule-based tools need ongoing maintenance. Someone must define every threshold and keep rules current as your product changes. Automated scoring reduces that ops load if you want scores updated continuously without manual spreadsheet work.