Review Customer Health

Build an evidence-backed view of whether a customer is achieving the outcomes they purchased for, what has materially changed, and what deserves attention next.

Use the organization's real customer-health framework when one exists. Do not convert incomplete product activity, isolated support events, or subjective impressions into a definitive churn or health label.

1. Define health for the account

Establish:

• account and relevant business unit;
• products, plans, and accepted scope;
• lifecycle stage;
• customer's business goals;
• accepted success measures;
• key stakeholders and owners;
• review period;
• upcoming customer or commercial decisions;
• accepted risk definitions;
• accepted opportunity definitions;
• intended audience for the review.

Determine whether the output is intended for:

• the account owner;
• a Customer Success team review;
• leadership;
• a portfolio review;
• renewal or expansion planning;
• an upcoming customer conversation.

Use the customer's accepted success plan and previous approved account review when available.

Keep organization-wide health rules separate from customer-specific commitments, expectations, and success criteria.

2. Build the current evidence picture

Use permitted, relevant sources such as:

• CRM and account records;
• meaningful product usage and adoption data;
• support history;
• onboarding and implementation progress;
• meetings and transcripts;
• relevant email;
• team communication;
• billing context;
• renewal context;
• open commitments;
• accepted account plans.

For consequential signals, preserve:

• source;
• date;
• applicable period;
• definition;
• coverage;
• owner;
• important limitations.

Consider signals that materially affect this account.

Outcome progress

Assess progress toward the customer's stated goals and accepted success measures.

Adoption and product value

Evaluate meaningful use connected to the customer's workflow and intended outcome.

Do not equate login counts, event volume, or raw activity with customer value unless the accepted model explicitly does so.

Support experience

Consider:

• unresolved customer impact;
• recurring issues;
• severity under the team's actual policy;
• workaround quality;
• resolution experience;
• outstanding escalations.

Stakeholder relationships

Consider:

• engagement of relevant stakeholders;
• champion strength where supported;
• sponsor involvement;
• stakeholder changes;
• decision-maker access;
• relationship gaps.

Do not infer relationship risk solely from message frequency.

Delivery and commitments

Track onboarding, implementation, product, support, or service commitments and whether they are being met.

Commercial context

Where approved and relevant, consider:

• renewal timing;
• expansion discussions;
• contraction context;
• approved commercial dependencies.

External change

Include known organizational, strategic, operational, or market changes only when they have a supported connection to customer success.

3. Reconcile signals instead of averaging away complexity

Compare signals in context.

Strong product use can coexist with an unresolved critical issue. Low activity can be expected for a periodic workflow. A quiet stakeholder may reflect timing rather than disengagement.

Separate:

• observed facts;
• customer statements;
• team judgment;
• analytical inference;
• stale information;
• unknowns.

Identify contradictions explicitly.

If the organization uses a health score, apply only its accepted:

• components;
• definitions;
• thresholds;
• weights;
• calculation logic.

Show the evidence behind material components.

Do not invent weights, thresholds, probabilities, churn percentages, or expansion likelihood.

Never infer protected or sensitive personal traits.

When the model cannot be applied reliably because data is missing or definitions conflict, report the limitation rather than manufacturing a score.

4. Identify material changes

Compare the current state with the most relevant prior accepted view.

Highlight changes that may affect:

• customer outcomes;
• adoption;
• relationships;
• delivery;
• support experience;
• renewal readiness;
• expansion readiness;
• risk;
• next actions.

Distinguish a real change from:

• reporting noise;
• incomplete coverage;
• seasonal behavior;
• definition changes;
• delayed data.

5. Deliver the account review

Lead with:

1. current outcome/value position;
2. material changes;
3. most important next decision or action.

Include as useful:

• progress against goals and success measures;
• supported strengths;
• supported risks;
• supported opportunities;
• meaningful adoption changes;
• stakeholder and relationship changes;
• open commitments;
• blockers;
• owners;
• accepted dates;
• missing evidence;
• conflicting evidence;
• prioritized next actions.

For each important risk or opportunity, make the basis inspectable.

When the next task is a QBR, EBR, check-in, renewal conversation, or other customer meeting, route preparation through the appropriate meeting-preparation workflow.

When unresolved customer impact requires cross-functional ownership, route it through the customer-issue escalation workflow.

6. Keep actions and judgments controlled

Treat these as separate actions:

• preparing the health review;
• drafting an account plan;
• contacting the customer;
• changing CRM fields;
• changing an official health score;
• changing renewal status;
• offering commercial terms;
• making a product commitment;
• committing delivery work.

Follow active scoped permission for each action.

Approval of an analytical review does not automatically authorize changes to customer records, commercial treatment, communication, or commitments.

Verify completed system changes where possible.

7. Preserve the accepted review method

After the team confirms the sources, definitions, materiality, review behavior, and output structure, preserve them for future cycles.

A recurring health workflow may prepare a draft review from named sources and surface material changes.

It should stop and request review when:

• the account identity changes;
• goals or success measures materially change;
• the health model changes;
• source coverage becomes unreliable;
• permissions change;
• the intended audience changes;
• sensitive context appears outside the approved scope;
• definitions or thresholds conflict.

Recurring analysis should not silently alter official health classifications or customer records.