Risk detection and intervention

Set what triggers an investigation, who owns it and what evidence you will record. A sponsor change, falling product use or unresolved support issue may warrant attention, but each can have several explanations. Build a process that checks the signal, assigns an owner and records whether the intervention helped.

Turn signals into questions

Use a small number of indicators that relate to the customer's goals. Gainsight's workflow examples include usage drops, support cases and sponsor changes. Its scorecards can combine several measures, but a single colour can hide which measure changed. Treat a signal as a prompt for assessment, not a conclusion about customer risk.

Interpret product activity against the customer's expected use; a change in activity alone does not establish dissatisfaction.

Define what counts as a meaningful change

A scorecard can separate customer health into measures instead of compressing every signal into one judgement. Gainsight supports measures such as product usage, customer support, renewal chances and product bugs, and allows related measures to be grouped—for example, ticket age and ticket severity in a support group.

Set each measure's grading criteria before you rely on its score. Gainsight's example for a product-usage measure grades 0–50 as poor, 51–75 as average and above 76 as very good; those are example settings, not universal risk thresholds. Scores can use numbers, letters or colours, and measures can be scored manually or automatically through Rules Engine.

If customers need different scoring methods, Gainsight allows multiple scorecards to be maintained, but only one can be assigned to a customer or relationship at a time. An administrator configures measures, groups, grading schemes, weights, exceptions and validity periods; users can view company scores on C360 and relationship scores on R360.

Customer Health Scorecard Grading Criteria (Gainsight Example)

For deeper usage analysis

For detailed guidance on investigating declining adoption, see the supporting article on that topic.

Intervene with a named owner

Record the concern, evidence, customer impact, next action, owner and review date. Customer success can coordinate an investigation and explain the plan, but a defect may require product or engineering, and a commercial issue may require leadership. Give the customer a realistic update rather than promising a fix another team has not approved.

Close a risk only when the underlying issue is resolved or there is a documented reason to reclassify it. A completed check-in is an activity, not evidence of recovery. Track the customer's outcome and whether the relevant behaviour returns over a suitable period.

Make the response trackable

Gainsight Cockpit uses Calls to Action (CTAs) as customer-linked action items. A CTA can be created manually or triggered automatically by business rules, including from drop-in usage, support cases, sponsor changes or survey results. A CTA typically contains tasks for the customer success manager to complete, making the response a sequence of work rather than just a health rating.

Use the Cockpit List View to group and sort CTAs, or Task Views to organise work at task level. Users can log and access Timeline activities from a CTA's detail view. These features make action items and updates visible to others in the organisation, and CTAs can link to opportunities, activities, contacts and cases.

Agree the CTA's type and priority with the team's configured workflow. Gainsight provides Risk, Expansion, Lifecycle, Activity and Objective CTA types; administrators can modify the names of several default types or create custom types. Before the team creates or manages CTAs, administrators must configure settings such as types, reasons, priorities and standard layouts.

Use reporting to review whether the work is progressing, not merely whether a task was opened. Gainsight describes CTAs as reportable action items for communicating progress to the team, and business rules can trigger them based on customer data or time factors. A rule can surface a signal for investigation.