Assessing Customer Discovery Skills: The 30-Minute Role-Play Design
Updated · Tech checked
Run a role-play where the interviewer plays a customer with a vague problem and a hidden constraint; score the questions asked, listening behavior, constraint discovery, and the candidate's restraint from premature solutioning.
The short answer
You can't assess discovery with questions about discovery - you need a live role-play. 30 minutes, one scenario, one scorecard. It's the highest-signal 30 minutes in the FDE loop.
Scenario design
- Vague opener: "Our order processing is a mess, we think we need AI." (Real customers say this.)
- Hidden constraint 1: the legacy system has no API - only nightly CSV exports. Surfaces only if asked about systems.
- Hidden constraint 2: a compliance rule that payment data can't leave the EU region. Surfaces only if asked about data handling.
- A planted number: "we process about 4,000 orders a day" - strong candidates anchor math on it; weak ones ignore it.
- Pressure moment: at minute 20, the "customer" says "so can you just build the AI part this sprint?" - watch the pushback.
Scoring sheet (1-4 each)
| Behavior | Signal |
|---|---|
| Process questions before solutions | core skill |
| Quantifies impact (uses 4,000/day) | business fluency |
| Discovers CSV-only constraint | system curiosity |
| Discovers data-residency constraint | risk instinct |
| Defines success metric in customer terms | outcome orientation |
| Handles the sprint pressure without overcommitting | boundary maturity |
Red flags: lecturing about their past stack; asking zero questions in the first 10 minutes; promising the sprint.
After the role-play
Ask: "What would you do first if this were real?" - a 60-second scoping summary tests whether they listened (it should reference the constraints they found). Calibrate interviewers by running one debrief together; the hiring guide integrates this stage into the full loop.
Continue: Discovery question bank · Hiring playbook