Customer Health Score
A customer health score is a composite indicator, often expressed from 0 to 100, that combines signals such as usage, satisfaction and engagement to estimate how likely a customer is to stay and grow.
Definition
A customer health score gives each customer account a single reading, often green, orange or red, so a customer success team knows where to look first. It combines several signals into one number: how much the product is used, how satisfied the customer is, how often they contact you, how they pay, and how engaged their sponsors are.
There is no standard formula, so you build your own. The usual method is to pick four to six signals, give each a weight according to how well it predicts departures, score each from 0 to 100, and add the weighted results. Then compare the score with what actually happened to past customers and adjust the weights. A score that does not match reality is worse than no score.
The strength of a health score is that it triggers action before a customer announces they are leaving. Its weakness is false confidence: a green score built on usage alone can miss an unhappy customer. Complete the figures with conversations, because customers are the only ones who can tell you what the numbers cannot. And remember that it is a tool for people, who decide what to do.
Key points
- It combines signals such as usage, satisfaction and engagement into one reading per customer.
- There is no standard formula: choose signals and weights, then check them against past outcomes.
- Show it as a simple score or a colour so that teams know which accounts to look at first.
- Complete it with conversations, because a number cannot explain a customer's real situation.
- Review the model regularly: products and customers change, and so do the warning signs.
Example
A team scores accounts with four signals: usage (weight 40%), support experience (20%), satisfaction (20%) and engagement in meetings (20%). One customer scores 80 on usage, 60 on support, 70 on satisfaction and 50 on engagement. Health = 0.4 x 80 + 0.2 x 60 + 0.2 x 70 + 0.2 x 50 = 32 + 12 + 14 + 10 = 68, which the team classes as orange and follows up with a call.
Frequently asked questions
How do I choose the signals of a health score?
Start from what you already know about customers who left and customers who grew: what did they do differently in the months before? Pick a handful of signals that you can measure reliably, and avoid adding more until the first version proves useful.
Is a health score the same as NPS?
No. NPS is a survey answer from the customer, while a health score is a composite that you build from several sources. NPS or CSAT results can be one input among others, next to usage and engagement.
How often should the score be updated?
As often as the underlying signals change and as often as the team can act on it. Many teams refresh it weekly or monthly. A score that nobody looks at or that is out of date does not help to anticipate churn.
Related terms
AI conversations, human decisions
Passisto holds structured voice conversations with candidates, employees, clients and students, and turns them into reports your team can act on.