Most SaaS teams already know when a customer is unhappy, but they learn it too late, from a cancellation message. The warning signs were there all along: logins faded, a key contact stopped replying, support tickets went unanswered, or the team never adopted the feature they bought the product for. A customer health score gathers those scattered signals into one simple view, so your team can act while there is still time to help.
A health score does not need complex software or dozens of inputs. In fact, a simple score that your team understands and trusts beats a sophisticated model that nobody checks. The goal is a shared early-warning system, not a perfect prediction.
Pick a Few Signals That Reflect Real Value
Choose three to five signals that show whether a customer is getting value. Useful signals often include how often the team uses the core feature, whether usage is rising or falling, how quickly support questions are resolved, and whether the main contact is still engaged. Avoid signals that look impressive but mean little, such as total page views, which can rise while real value falls.
For each signal, define what good, fair, and poor look like in plain terms. For example, regular weekly use of the core feature might count as good, occasional use as fair, and no use for several weeks as poor. Clear definitions keep scoring consistent when different team members review different accounts.
Weight Simply and Review Together
Give each signal a simple weight based on how strongly it relates to staying or leaving. Usage of the core feature usually deserves the heaviest weight, because a team that stops using the product has already started leaving in practice. Support experience and contact engagement matter too, but they rarely outweigh sustained non-use. Keep the maths visible: anyone on the team should be able to explain why an account scored as it did.
Review scores together on a regular rhythm, such as weekly for at-risk accounts and monthly for the full list. Use the review to assign one owner and one next action per at-risk account, such as a check-in call, a training offer, or a review of their setup.
Act Early, and Improve the Score Over Time
The value of a health score lies in early action. When an account slips, reach out with curiosity rather than alarm: ask what changed, offer help with the specific problem, and record what you learn. Over time, compare scores with what actually happened. If accounts you marked healthy still left, your signals are missing something. If accounts you flagged recovered after a simple intervention, your system is working.
A health score that evolves with real experience stays trusted, while a frozen score slowly becomes decoration. Start simple, act early, and let the score earn its keep. For related guides, see How to Write SaaS Release Notes Your Customers Will Actually Read and How to Run a SaaS Renewal Review That Keeps Customers Longer.