92. An experiment improves overall task completion but makes a keyboard-only path unusable. How would you present the result?
HardEvaluation
Case
Show model answer
Report the aggregate improvement and the blocked path explicitly, including the affected scope and relevant obligations. The average does not cancel an accessibility barrier, and limited event data may undercount people who abandon early.
Work with accessibility, product and engineering owners on mitigation and release decisions. Validate the corrected path with appropriate methods instead of concluding that a statistically favorable overall result proves the experience is acceptable.
What interviewers listen for
- Expose the blocked path alongside aggregates
- Assess scope and applicable obligations
- Validate mitigation with relevant users