Customer health score: which customer is at risk?

Most companies learn a customer is leaving only when that customer hits the cancel button — that is, after it's too late. Yet customers usually drift away not in a day but over weeks: usage drops, emails go unanswered, support tickets rise or they go entirely silent. All these signals are there; the problem is nobody looks at them from one place. The customer health score solves exactly this: it gives each customer a single composite score (e.g. 0–100) that shows their trajectory in advance.

This score lets you see whether a customer is "healthy" (happy, open to growth) or "at risk" (drifting, near cancellation) before they tell you — so you step in before churn.

This article covers what a customer health score is, how it differs from the deal health score and churn, which signals it's made of, how to build it, what to do with the score, and how to track it with a CRM.

Customer health scoreAt riskHealthy

What is a customer health score?

A customer health score is a single number that sums up how healthy a customer's (account's) relationship with you is. It gathers multiple signals (usage, engagement, support, payment, relationship) into one composite score and usually ties it to a color code (green / yellow / red). Its purpose is predictive: not to report the past but to signal the future ahead of time — especially a customer's risk of drifting away. This score is the early-warning system of the post-sale world; the compass of a customer success team.

How does it differ from the deal health score and churn?

This concept is often confused with its two neighbors but is different from both. The deal health score measures a deal's (a not-yet-sold opportunity's) probability of closing — so it's pre-sale. The customer health score measures an existing customer's probability of staying/growing — so it's post-sale. Churn prevention is a strategy (what you'll do); the health score is the measurement tool of that strategy (telling you whom to intervene with, and when). In short: deal health answers "will this deal close?", churn prevention "how do I keep the customer?", and the health score "which customer is at risk?"

Which signals is it made of?

A solid health score rests not on one data point but on several dimensions. Usage: how often and how deeply does the customer use the product/service? Declining usage is the strongest early warning. Engagement: do they join your emails, calls, meetings, or have they gone silent? Support: are they having lots of problems, are their tickets resolved, how's their satisfaction (e.g. NPS)? Payment: are invoices on time, are there payment issues? Relationship: is there someone in the org who advocates for you, or has the only contact left? Together these dimensions give a balanced picture that a single signal can't mislead.

How do you build the score?

Building a health score doesn't have to be a big data-science project; it can start simple and transparent. First choose the most meaningful signals for your business (for most companies usage and engagement are the strongest). Then give each signal a weight — not all are equally important; declining usage may be a more critical warning than a late payment. Next define thresholds: which score range is green (healthy), which yellow (watch), which red (urgent intervention). At first you set these weights by intuition; over time you see which signal really predicts churn and improve the score. What matters is starting with a simple but actionable score.

What do you do with the score?

A health score is valuable only when it's tied to action. Red (at-risk) accounts should be prioritized and reached proactively: a call to understand the reason, training, resolving an issue or refreshing the relationship. This is the most effective way to stop churn before it happens. Green (healthy) accounts are an opportunity: these happy customers are the most open to upsell and cross-sell and the best referral sources. So the score guides both defense (churn prevention) and offense (growth) — directing scarce customer-success time to where it's needed most.

Tracking the health score with a CRM

The only place a customer health score can live is the CRM, where all of the customer's data sits. The CRM combines usage, engagement, support and payment signals into a single customer record and turns them into a score. When the score is visible in the customer list with a color code, the team can look every morning at "which accounts turned red today?" and set their priorities. The CRM can also trigger an automatic alert when the score drops — so an at-risk customer enters the team's radar without anyone waiting to notice. This score is the most direct way to protect a customer's long-term lifetime value and a core tool of a customer success team.

Health score and the customer lifecycle

What counts as a "healthy" customer changes with their stage in the lifecycle. For a new (onboarding) customer, health is measured by how fast they adopt the product — if seeing first value is delayed, risk rises. For a mature customer, health is measured by steady usage and expansion (new teams, new modules). For a customer nearing renewal, the engagement of recent months becomes critical. So a single fixed threshold doesn't fit all customers; a good health score takes into account which stage the customer is in. The same usage drop can mean "still getting used to it" in a new customer and "losing interest" in a year-old one.

Leading and lagging signals

Splitting signals in two gives a powerful view of the health score. Leading signals warn of the future in advance: declining usage, unanswered emails, falling session frequency. These give a chance to intervene before it's too late. Lagging signals show the outcome: a cancellation request, a payment stop, an open complaint. Most companies look only at lagging signals — but by that point the customer has already decided to leave. The real value of the health score is catching leading signals and warning while it's still early. When building your score, give the weight to leading signals; because your goal is not to report churn but to prevent it.

Health score and revenue: why it matters

The health score is not just a customer success tool but a direct revenue tool. Retaining an existing customer is many times cheaper than acquiring a new one; therefore rescuing a red account in time is concrete revenue protection. At the same time green accounts are the source of expansion (upsell, cross-sell, renewal) revenue. The health score makes both of these revenues visible: which revenue is at risk and which is ready to grow. In subscription-based businesses, this directly affects net revenue retention (NRR). In short, the health score lets you manage your customer base not like a cost center but like a revenue asset to be protected and grown.

Example: two customers, the same appearance

Both customers keep their subscription and pay their invoices; on the surface both look "fine." But the health score shows what's behind the curtain: the first customer's usage has halved in the last two months, they haven't replied to the last three emails and their only contact has left the company — their score is red. The second uses it regularly every week, joins calls and has added a new department to the system — their score is green.

Without a health score, the team would see both equally as "a paying customer" and neglect them; the first would quietly cancel renewal, and the loss would be noticed only when the invoice stops. With the score, the team reaches the red account in time and prevents churn, and brings an upsell offer to the green account. The same two customers — the difference is making risk visible and acting at the right moment.

Common mistakes

The four most common health-score mistakes: First, relying on a single signal — looking only at usage or only at NPS misleads; health is multidimensional. Second, not tying the score to action — a red score nobody intervenes with is just an observation. Third, focusing only on reds and forgetting greens — healthy customers are the biggest growth (upsell, referral) opportunity. Fourth, never updating the score — if you don't learn which signal really predicts churn and improve the weights over time, the score goes dull.

Summary: where to start

The customer health score turns customer loss from a surprise into something predictable. First choose the strongest signals for your business (usually usage and engagement), define a weight and green/yellow/red thresholds for each, calculate the score in the CRM together with all the customer data and, most importantly, tie it to action: intervention for red, a growth offer for green. Over time learn which signal really predicts churn and improve the score. With this approach you rescue customers before they cancel, grow the happy ones and systematically protect the value of your customer base.