Quick answer
A customer health score is a composite number, usually 0–100 or green/amber/red, that combines product usage, customer outcomes, relationship and commercial signals to estimate how likely an account is to renew, expand or churn.
Formula: Health score = Σ (weight × normalised input score)
What is a customer health score?
A customer health score turns many signals about an account into one number (often 0–100, shown as green, amber or red) that estimates whether the customer will renew, grow or leave. Customer success teams use it to decide where to spend time before a renewal is at risk.
Common inputs
| Group | Example inputs |
|---|---|
| Product usage | Share of licensed seats active weekly, use of key features, login trend |
| Outcomes | Goals reached, reports viewed by the buyer, value milestones |
| Relationship | Open support tickets, survey sentiment, executive sponsor engaged |
| Commercial | Days to renewal, failed payments, recent downgrade |
Usage inputs are usually the strongest predictors for software, because they move weeks before a cancellation.
A sample weighted scorecard
Each input is scored 0–100 first, then weighted (sample numbers):
| Input | Weight | Account A | Account B | Account C |
|---|---|---|---|---|
| Weekly active seats | 35% | 90 | 60 | 30 |
| Key feature usage | 25% | 80 | 70 | 20 |
| Support sentiment | 20% | 70 | 40 | 50 |
| Commercial (renewal, payments) | 20% | 90 | 50 | 40 |
| Health score | 83.5 | 56.5 | 33.5 |
Account C needs attention now. Account B is amber mainly because of support sentiment; its usage is fine.
Validate it before you trust it
A health score is a prediction, so test it:
- Score every account as it was at a past date, using only data available then.
- Look at what happened in the following quarter (sample numbers):
| Score band | Accounts | Churned next quarter |
|---|---|---|
| Red (< 40) | 60 | 30% |
| Amber (40–70) | 140 | 11% |
| Green (> 70) | 300 | 4% |
- If churn does not fall clearly from red to green, change the inputs or weights. If it does, check which single input separates them best; it may deserve more weight.
Pitfalls
- Size bias. Large accounts have more users and more activity; use rates (share of seats active), not totals.
- Too many inputs. Fifteen weighted signals are hard to explain and to act on. Five good ones usually do better.
- Lagging only. NPS once a year and renewal dates warn too late; include usage trends.
- No action attached. A red score should trigger a defined playbook, otherwise it is just a colour.
Over time a working score shows up in gross revenue retention and lower churn.
Usage inputs from TapCub
TapCub does not compute a health score, but it supplies the usage half. With account-level identity, the Accounts page shows activity per company; retention and lifecycle show whether an account's users keep coming back; and rule segments such as "accounts with no key-feature use in 14 days" can notify your customer success tool through a webhook when membership changes. For order-based businesses, an RFM segment buckets customers by recency, frequency and spend.
See also: Product intelligence.
See also
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.
Data Engineer
Writes about event design, data quality, and collection workflows, with verifiable implementation methods for reliable analytics.