One number, many signals
A customer health score rolls usage, engagement, support, and relationship signals into a single read on whether an account will renew, expand, or churn. Its job is to give customer success a fast, consistent way to triage a large base without deeply reviewing every account. Done well, it directs limited attention to the accounts that need it. Done poorly, it becomes a comforting dashboard color that has drifted from reality. The difference is entirely in the signals behind it and the action it triggers.What separates a good score from a vanity one
| Reliable health score | Vanity health score |
|---|---|
| Signals that predict outcomes | Inputs chosen because they are easy |
| Weights validated against churn and renewal | Weights guessed once and frozen |
| Updated as data teaches | Static, drifts from reality |
| Triggers specific action | Sits as a color on a dashboard |
The score is a trigger, not a label
A health score matters only if a red account produces an intervention while the outcome can still change. Used to prioritize outreach and catch decline early, it is a genuine retention tool that protects net revenue retention. Used as a passive label, it describes churn without preventing it. This is also where AI churn prediction strengthens the score, weighing many signals more consistently than a hand-tuned formula, provided the underlying usage and CRM data is clean. The number is the start of the work, not the end of it: health scoring reduces churn only when a declining score reliably becomes a phone call.
What should a customer health score include?
Choose signals because they predict renewal and churn, and leave out the ones chosen because they were easy to pull. These are the signals ORM finds most useful:
| Signal | What to measure | What healthy looks like |
|---|---|---|
| Support pattern | Tickets in the last 12 months, and their tier | Three to five routine cases a year |
| Engagement | Days since meaningful activity: logins, meetings, replies | Regular contact with no long silences |
| Product usage | Adoption of the features the customer bought | Stable or rising |
| Contraction | Dropped products or reduced seats | None |
| Renewal timing | Days to the renewal date | Enough runway to act if other signals turn |
How do you calculate a customer health score?
Score each signal on the same scale, weight them, and add them up. Here is a hypothetical example. The weights are illustrative, so validate your own against which accounts actually renewed.
| Signal | Weight | Account score (0 to 100) | Weighted |
|---|---|---|---|
| Product usage | 30% | 80 | 24.0 |
| Support pattern | 25% | 30 (no tickets in a year) | 7.5 |
| Engagement | 25% | 60 | 15.0 |
| Relationship and contract | 20% | 70 | 14.0 |
| Health score | 60.5 |
Why do most health scores read support tickets backwards?
Because they treat tickets as a cost. Most scores subtract points as tickets rise, which rewards the customer who never calls. In ORM data the relationship is a U-curve:
| Tickets in the last 12 months | What it usually means | How to score it |
|---|---|---|
| None | The customer is not using the product deeply enough to hit friction | Churn risk |
| Three to five, mostly tier 2 or tier 3 | Engaged, getting help and getting value | Healthiest |
| Seven or more | Persistent problems | Churn risk |
Frequently Asked Questions
How do you calculate a customer health score?
Score each signal on the same scale, such as 0 to 100, then combine them with weights. Validate the weights against which accounts actually renewed or churned, and adjust them as the data teaches you. A score whose weights were guessed once and never checked will drift.
Do more support tickets mean a customer is unhealthy?
Not in a straight line. In ORM data the pattern is U-shaped. No tickets in a year is a churn risk, and so is seven or more. Three to five, usually routine tier 2 or tier 3 cases, is the healthiest profile.
What is a customer health score?
It is a composite indicator that rolls up signals like product usage, engagement, support history, and relationship strength into one score reflecting an account's likelihood to renew, expand, or churn. The point is to give customer success a fast, consistent read on which accounts are healthy and which need attention, without reviewing every account in depth.
What makes a health score reliable?
Signals that genuinely predict outcomes, weighted by how much they actually correlate with renewal and churn, and kept current. A score built on vanity inputs or fixed guesses about weighting drifts from reality and trains the team to ignore it. The best scores are validated against actual outcomes and adjusted as the data teaches which signals matter.
How do you act on a health score?
Route declining accounts to intervention early and use the score to prioritize limited customer success attention, not as a label. A score is only valuable if a red account triggers a specific action while the outcome can still change. Treated as a dashboard color with no follow-through, it measures risk without reducing it.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like customer health score into prescriptive action for your team.
Schedule a Demo