Skip to content

Latest commit

 

History

History
40 lines (31 loc) · 1.43 KB

File metadata and controls

40 lines (31 loc) · 1.43 KB

User-research guide

Do not demo first. Ask about the last actual acceptance event.

Participants

  • 3 daily operators: QA, accessibility, security, delivery, or IV&V
  • 3 people near authority or budget: product executive, COR, procurement, prime
  • 2 potential partners: testing firm, public-interest digital service vendor
  • 2 skeptics who believe current tools already solve the problem

Core questions

  1. Show me how the last software deliverable was accepted.
  2. Where was the authoritative requirement?
  3. Who decided what evidence would prove it?
  4. Which evidence was generated automatically, reviewed manually, or supplied by the vendor?
  5. What was impossible to verify?
  6. How did you distinguish missing evidence from a failed requirement?
  7. What happened when the source requirement changed?
  8. Could a reviewer reproduce the acceptance package six months later?
  9. Which existing system was supposed to do this?
  10. What did the gap cost in delay, rework, dispute, or risk?
  11. Who could fund a bounded improvement?
  12. What would make this generic consulting rather than a product?

Mapping exercise

Give two experts the same ten-clause public excerpt. Independently record:

  • classification;
  • required evidence;
  • blocking criticality;
  • accountable owner;
  • ambiguity;
  • whether another tool already owns the problem.

Measure agreement before reconciling. Disagreement is evidence about the product, not an evaluator error to hide.