Customer health score examples are useful only when they fit the customer motion. Copying a template is easy. Trusting it is harder.
Imagine a Customer Success leader drops one neat spreadsheet into the weekly account review. Every customer gets the same inputs: product usage, support tickets, survey score, last meeting date and renewal timing.
At first, the model looks sensible. Then the edge cases arrive. A new onboarding customer is red because usage is low, although implementation is on track. A self-serve account is green because logins are rising, while payment has failed twice. A strategic enterprise account is amber because the champion is positive, even though the executive sponsor has disappeared.
The problem is not the spreadsheet. The problem is that one model is being asked to answer too many questions.
The best customer health score examples show the structure of a decision. They do not prescribe universal weights. Use the examples below as illustrative starting points, not benchmarks, best-practice weightings or statistically validated recommendations. The practical principle is simple: copy the structure, not the weights.
What every customer health score template should include
A useful customer health score template separates five layers: raw evidence, scored input, overall band, confidence and next action.
Raw evidence is what happened. A scored input is how the team interprets it. The overall band summarises the current view. Confidence tells the team whether that summary is trustworthy. The next action turns the score into work.
This separation matters because exact-looking scores can hide weak data. The UK Government Data Quality Framework describes data quality as fitness for purpose and highlights completeness, consistency, timeliness, validity and accuracy (UK Government Data Quality Framework). In a health score, a stale usage feed, missing survey response or undefined support measure should not become green by default.
Public health-scoring examples support the same caution. GitLab combines lenses such as product, risk, outcomes, voice of customer and engagement, and makes stale or missing measures visible (GitLab Handbook). Gainsight describes scorecards as measures, measure groups, grading schemes, weights, exceptions and validity periods, with multiple scorecards where one method does not apply to every customer (Gainsight Support).
| Field | Why it matters |
|---|---|
| Segment or lifecycle stage | Prevents one model being forced across different customer motions. |
| Decision the score supports | Keeps the score tied to prioritisation, escalation, onboarding help or renewal review. |
| Inputs | Shows the evidence included. |
| Weight or rule | Explains how the input affects the score. |
| Freshness and confidence | Stops stale, missing or low-quality data looking precise. |
| Owner | Makes someone accountable for review and action. |
| Triggered next action | Turns the score into work rather than reporting. |
| Review cadence | Keeps the template from decaying. |
Example 1: Simple spreadsheet health score for an early-stage CS team
Purpose: give a small team a first customer health score spreadsheet before automation or advanced modelling.
For an early-stage team, red, amber and green can be more credible than a 100-point score. If data is immature, the template should expose judgement, not disguise it.
| Illustrative input | Example weight or rule | Caution and action |
|---|---|---|
| Recent meaningful CSM contact | 20% | Define "meaningful". If evidence is stale, the CSM checks notes and follows up. |
| Onboarding or adoption milestone | 20% | Score only milestones that are due. If overdue, name the blocker. |
| Open support issue severity | 20% | Define severity consistently. Severe open issues need an owner. |
| Stakeholder responsiveness | 15% | Silence may have context. Confirm the right contact before scoring it as risk. |
| Renewal proximity | 15% | Timing changes urgency, not health by itself. |
| CSM judgement with written reason | 10% | Require a reason, confidence level and review date. |
Illustrative thresholds can stay simple: green means no known current risk and fresh evidence; amber means investigation is needed; red means an owner must act now. Missing evidence should reduce confidence. It should not make the customer green by default.
This model works because it is honest about its limits. It gives the team a shared language for review without pretending that an early spreadsheet is a mature customer health scoring model.
Example 2: PLG or self-serve SaaS health score template
Purpose: identify accounts where adoption, payment or support risk deserves attention at scale.
A product-led growth or self-serve model may not have regular CSM contact for every account. Product behaviour and commercial signals often arrive first, but they still need interpretation.
Amplitude's product adoption guidance discusses activation, feature usage and adoption rate, with activation tied to a predefined meaningful action (Amplitude product adoption metrics). A login is rarely enough. The question is whether the customer is doing the thing that creates value in that segment.
| Illustrative input | Example weight or rule | Caution and action |
|---|---|---|
| Activation completion | 25% | Define the activation event by segment. |
| Meaningful feature use | 25% | Avoid vanity usage. Tie use to the customer's value path. |
| Adoption breadth | 15% | Normalise by account size or expected usage rhythm where possible. |
| Active-user trend | 15% | Check seasonality, project cycles and tracking gaps. |
| Billing or payment status | 10% | Treat payment issues as commercial risk, not product sentiment. |
| Support friction or survey response | 10% | Low survey coverage should lower score confidence. |
An illustrative threshold might be green above 75, amber from 50 to 75 and red below 50, but only with current data. If activation data is missing for half the segment, show low confidence rather than presenting the score as precise.
For self-serve accounts, the score is usually a prioritisation tool. It helps the team decide where to inspect, nudge, route or review. It should not turn every dip in usage into a customer intervention without checking the evidence.
Example 3: High-touch enterprise health score template
Purpose: help CSMs and leaders understand strategic-account confidence before executive review or renewal planning.
Enterprise health is not just product activity. A large account may use the product heavily while political support weakens. Another may show modest usage but strong outcome progress because the rollout is narrow and deliberate.
| Illustrative input | Example weight or rule | Caution and action |
|---|---|---|
| Agreed outcome progress | 25% | Require evidence, not general optimism. |
| Executive sponsor strength | 15% | A friendly champion is not always an executive sponsor. |
| Stakeholder coverage | 15% | Check role coverage, not just contact count. |
| Product adoption depth | 15% | Link adoption to the enterprise use case. |
| Open escalations | 10% | Define escalation age, severity and owner. |
| Sentiment and voice of customer | 10% | Sentiment is one input, not the score. |
| Commercial and renewal confidence | 10% | High account value should affect priority, not automatically health. |
For enterprise accounts, missing sponsor evidence should appear as "unknown" or low confidence. Do not let usage keep the whole account green if the commercial relationship is single-threaded, outcome evidence is thin or escalation ownership is unclear.
This is where a health score should help the account team have a better conversation. It should make the weak evidence visible before the renewal or executive review, not after.
Example 4: Onboarding health score template
Purpose: identify implementation risk before a customer fails to reach first value.
An onboarding health score asks whether the customer is moving towards first value at the expected pace. It is not a renewal score.
Amplitude describes time to value as the moment when users solve the problem that brought them to the product, and recommends considering the path to value by segment (Amplitude time to value). In a template, ask: has this customer reached the agreed early value milestone, and if not, why not?
| Illustrative input | Example weight or rule | Caution and action |
|---|---|---|
| Implementation milestone progress | 25% | Compare with the customer's agreed plan, not a generic timeline. |
| First value status | 20% | Define first value before scoring it. |
| Blocker status | 20% | Name the owner and dependency. |
| Training or enablement completion | 10% | Completion does not prove adoption. |
| Stakeholder responsiveness | 10% | Non-response needs context before judgement. |
| Data or technical dependency | 10% | Keep dependencies generic unless approved source systems are confirmed. |
| Early support friction | 5% | Early tickets can mean setup activity or real friction. |
This template should expire or convert after onboarding. When the account moves into adoption, renewal planning or expansion review, use a different view. Otherwise the team may keep judging a mature customer by implementation logic long after implementation has ended.
Example 5: Support-led customer risk score template
Purpose: detect customer experience risk where support friction is a major part of health.
Support data can show where the customer is blocked. It can also mislead. High ticket volume may signal poor experience, active implementation, complex rollout or advanced usage.
Definitions matter. Zendesk defines first reply time as the time between ticket creation and the first public agent comment, and notes that reporting may distinguish calendar hours, business hours and channel behaviour (Zendesk first reply time). A support-led score needs the same clarity.
| Illustrative input | Example weight or rule | Caution and action |
|---|---|---|
| Unresolved severity | 25% | Severity must be applied consistently. |
| Repeat issue themes | 20% | Group themes carefully; do not count duplicates blindly. |
| First reply or resolution experience | 15% | State whether the measure uses business or calendar hours. |
| Escalation age | 15% | Age alone is not enough without severity. |
| Customer sentiment | 10% | Qualtrics cautions that NPS should not be treated in isolation (Qualtrics NPS). |
| Recent usage movement | 10% | Check whether a support issue blocked value. |
| Owner follow-up | 5% | A logged task is not the same as completed follow-up. |
This template should not turn every support-heavy customer red. Ask whether support friction is affecting value, trust or renewal confidence. A customer with many low-risk implementation questions needs a different response from a customer with one unresolved severe issue blocking work.
Example 6: Renewal-readiness health score template
Purpose: prepare the team for retention and renewal conversations.
A renewal-readiness score is not a churn formula. ChartMogul defines customer churn as customers leaving through subscription cancellations (ChartMogul customer churn). A renewal-readiness template should focus on current evidence the team can still act on before the commercial moment.
| Illustrative input | Example weight or rule | Caution and action |
|---|---|---|
| Realised outcomes | 25% | Outcomes need proof, not just activity. |
| Adoption trend | 15% | The trend must fit product rhythm. |
| Relationship strength | 15% | Single-threaded accounts are fragile. |
| Open risks | 15% | Risks need owner, severity and due date. |
| Commercial blockers | 10% | A blocker is not always product dissatisfaction. |
| Renewal date proximity | 10% | Proximity changes urgency, not value. |
| Support escalations | 5% | Align support and renewal communication. |
| Expansion or contraction signals | 5% | Treat as context, not a guaranteed result. |
Use this template early enough to change the conversation. If renewal is next week, the score may prioritise attention, but it will not repair missing outcome evidence.
For a renewal-readiness view, the score should also make ownership obvious. A red commercial blocker, a red product escalation and a red relationship signal may all matter, but they usually need different owners.
How to adapt these templates without copying the weights
Use the examples as scaffolding. The weights are illustrative starting points, not a standard.
- Choose the segment and lifecycle stage.
- Define the decision the score should support.
- Pick the few inputs that provide timely evidence.
- Define each input before scoring it.
- Set illustrative weights and thresholds for your first review cycle.
- Add freshness and confidence rules.
- Decide the owner and action for each band or trigger.
- Review false positives, missed risks and stale inputs after use.
The most important adaptation is often subtraction. If a signal does not change a decision, remove it. A short template a CSM can explain beats a long model no one believes.
Also decide what the score is not for. A model built for CSM prioritisation may not be reliable enough for renewal forecasting. A model built for onboarding intervention may not be useful for board reporting. One customer health score can support shared language, but it should not be forced to serve every operational decision.
Common mistakes with customer health score examples
Most mistakes are operating mistakes, not mathematical ones.
- Copying weights from another company.
- Mixing onboarding, adoption, enterprise relationship and renewal decisions in one score.
- Treating missing data as healthy.
- Over-weighting vanity usage such as logins.
- Letting subjective judgement hide the evidence.
- Adding too many inputs for the team to maintain.
- Using one score for leadership reporting, CSM prioritisation and renewal forecasting.
- Failing to record why the score changed.
- Triggering outreach before checking the evidence.
Here is an illustrative before-and-after comparison.
| Generic model | Why it misleads | Better segment-specific version |
|---|---|---|
| Product usage: 40% for every customer | Usage means different things in onboarding, self-serve and enterprise accounts. | Define meaningful usage by segment and lifecycle stage. |
| Support tickets: green when ticket count is low | Low ticket count may mean no issues, no engagement or poor reporting. | Score severe unresolved issues, repeat themes and support experience. |
| Sentiment: green if champion is positive | A happy champion can coexist with weak sponsorship or budget pressure. | Treat sentiment as one input beside stakeholder and commercial evidence. |
| Missing fields ignored | The score looks precise even when evidence is incomplete. | Show confidence separately and flag missing or stale inputs. |
| Red score triggers automatic outreach | The team may contact the customer before checking the alert. | Require evidence review and owner assignment before customer action. |
The better version is more honest about what the team knows. That honesty is what makes the score useful. It lets the team distinguish "this customer is unhealthy" from "we do not yet have enough evidence to trust the score".
A blank customer health score template to adapt
Use this blank customer health scorecard template for one segment or lifecycle stage at a time.
| Input | Definition | Source | Direction | Example weight | Freshness rule | Confidence | Owner | Triggered action |
|---|---|---|---|---|---|---|---|---|
| [Metric] | [What exactly counts] | [System or person] | Positive or negative | [Illustrative only] | [Recency] | High, medium or low | [Role] | [Next step] |
| [Metric] | [What exactly counts] | [System or person] | Positive or negative | [Illustrative only] | [Recency] | High, medium or low | [Role] | [Next step] |
| [Metric] | [What exactly counts] | [System or person] | Positive or negative | [Illustrative only] | [Recency] | High, medium or low | [Role] | [Next step] |
| [Metric] | [What exactly counts] | [System or person] | Positive or negative | [Illustrative only] | [Recency] | High, medium or low | [Role] | [Next step] |
Before using it, add four fields above the table: segment, lifecycle stage, decision supported and review cadence. Then add one field below it: what would make this score untrustworthy?
That last question prevents the worst failure mode. When evidence is missing, stale or poorly defined, the template should say so.
The right customer health score template will not remove judgement from Customer Success work. It gives that judgement a cleaner frame: which evidence matters, how current it is, who owns the response and what should happen next. Start with the example that matches the decision in front of you, adapt the inputs to your customer motion, and keep confidence visible alongside the score.