How do I build a customer health score?

Build a customer health score by combining the signals that predict whether an account will stay, grow, or leave (product usage, support experience, billing, and relationship) into one weighted, explainable score, then translate that number into a Health Tier written onto the account in Salesforce. Health is to retention what Fit is to acquisition, a single comparable measure that tells the team where to spend its time.

What sets this apart is where the signals come from. Where most tools stop at external public data, SignalRx also reads your own internal history, revenue, support, and product usage tracked over time, and scores both accounts and their contacts. When two sources disagree, inference decides which to trust. A decline shows up in that internal record weeks before the customer says anything, and the resulting scores and tiers become Revenue Plays and working lists native to Salesforce.

This guide covers how the score is built. For the companion breakdown of the individual warning signs that feed it, see What signals predict churn? For how the score drives action, see How do I catch at-risk customers?

What signals go into health?

The strongest inputs are usage (are they using what they pay for), support (the volume and sentiment of issues), billing (on-time payment, overages, downgrades), and relationship (engagement and whether a champion is present). Each is weighted by how well it predicts churn for your product, and SignalRx computes the blend from the revenue, usage, and support data you already have. The relationship signal is mostly your own internal engagement data, with a little external context to sharpen it: news and the web flag an M&A or ownership change at the account, and LinkedIn flags a champion's departure or a new decision-maker, both fed in as inputs to the relationship signal rather than replacing the internal read. The mix travels across healthcare, whether the account is a specialty group, a senior living operator, an FQHC, a behavioral health provider, a specialty pharmacy, a dental or aesthetics group, or an MSO or PE-backed platform, and where an account record is thin, SignalRx fills the gaps from public sources so every account still gets a score.

Illustrative example

Illustrative, based on how SignalRx customers work. A vendor weights its four signals to reflect that usage decline is its clearest predictor, then lets the score fall into tiers.

Should health be a score or a tier?

Both, the same way Fit carries a number and a set of tiers. A number powers dashboards and thresholds, and a Health Tier (Expand, Maintain, Triage, At Risk) is what a CSM reads at a glance. The tiers drive the alerts and the Call Now list, so the whole team acts off one shared label.

Why does explainability matter for health?

A CSM needs to know why an account is At Risk to do anything about it. SignalRx attaches the reasons (usage dropped, tickets spiked, a payment slipped) to the score, so the number is a starting point for the save conversation rather than a mystery to investigate.

Key takeaways

• Build health from usage, support, billing, and relationship signals.
• Keep both a number and a tier, weighted to predict churn for your product.
• External context (news and web for ownership change, LinkedIn for a champion move) complements the internal signals, it does not replace them.
• Attach the reasons so CSMs know exactly what to act on.

See it in action: Increase Customer Revenue with SignalRx.

Related: What signals predict churn? and How do I catch at-risk customers?