Which signals would you use to assess account health, and what would a health score fail to tell you?

Example Answer

I'd combine evidence about adoption, customer outcomes, engagement, unresolved issues and the commercial situation. I'd choose measures that fit the customer's stage and workflow: a new implementation and a mature account shouldn't automatically be judged by the same usage pattern. I'd check the source, definition and freshness of each signal before relying on it.

I'd use the score to help decide where to investigate, then read the underlying dimensions and recent account notes. A strong average can hide a serious blocker, and missing data can look reassuring if the scoring rules handle it poorly. I'd record the reason for concern and the action needed rather than just change a color.

The score can't tell me everything about budget, executive priorities or renewal intent. I'd confirm those through appropriate customer conversations and compare the model with actual outcomes over time. A healthy score is useful evidence, but it isn't a customer's commitment to renew.

Make it your own

Choose signals that fit the product, customer stage and account segment. Do not present a health-score threshold as universally predictive of renewal.

Why this works

Combines several kinds of evidence, checks their reliability and exposes risks that an aggregate score can conceal. The linked scorecard documentation describes configurable measures, not a proven universal prediction model.

Interviewer follow-up

What would you do if the score is green but the customer says they plan to leave?

I'd record the direct renewal risk and its source, then investigate the reason with the customer and account team. I wouldn't let the color override explicit evidence. I'd check whether the model missed a dimension, used stale inputs or weighted them poorly, and agree any model change with its owner rather than silently alter scoring rules.

References

• Gainsight's multidimensional scorecards and manual or automated measures
• GitLab's customer health assessment and risk coordination practices