How would you design a customer health score for a B2B SaaS product?
What the interviewer wants to hear
- Combine multiple signals: usage depth, adoption of key features, support, engagement, sentiment, outcomes
- Weight signals by how well they predict churn
- Segment-specific scoring
- Validate against historical churn and renewals
- Review and adjust regularly
Model answer
I would start by looking at accounts that churned and those that renewed or expanded, to see which signals actually separated them. Typical inputs are depth of product usage and adoption of the key features they bought for, number and severity of support tickets, engagement such as meeting attendance and executive contact, sentiment from NPS or CSM notes, and progress on agreed goals. I would weight each input based on how strongly it predicts churn, and I might use different models for SMB and enterprise because their behaviour differs. I would keep it simple enough that CSMs trust and understand it. Then I would test it against past data and review it every quarter, adjusting weights when a "green" account churns unexpectedly.
Can you answer it under time pressure?
Take a timed medium round and get scored on every answer.
More medium Customer Success Manager questions
- Walk me through how you prepare and run an effective QBR.
- Explain the difference between NRR and GRR, and why a company tracks both.
- How do you work with sales and product teams as a CSM?
- You have 150 accounts. How do you decide where to spend your time each week?
- A large customer bought 500 licences but only 150 are active after three months. How would you drive adoption?
- At kickoff, you realise sales promised something the product cannot do. How do you handle it?
- How do you identify expansion opportunities that genuinely help the customer?
- Walk me through how you would manage a major escalation from a key customer.
- How would you design a digital or tech-touch customer success programme for 2,000 small customers?