run-usability-testing

Run Usability Testing

Purpose

Help teams design and synthesize usability tests that surface friction, confusion, and failure points in product flows.

Skill type

Conceptual skill

Use this skill when

• A new design or flow is ready for testing before launch
• There are hypotheses about user confusion or friction in an existing flow
• A redesign needs to be validated against the original
• Post-launch feedback suggests usability problems

Do not use this skill when

• The goal is generative research (understanding needs) — use plan-ux-research
• No design or prototype exists yet

Required inputs

• Product flow or feature to be tested
• Key user tasks to test

Optional inputs

• Design prototype or staging environment
• Target participant criteria
• Hypotheses about friction points
• Prior usability test findings

Upstream context

Works best when:

• UX research plan exists
• Prototype or staging environment is available

If upstream context is missing

If no prototype exists, produce a usability test plan for when it is ready, with placeholder tasks.

Downstream handoff

Output can feed:

• assess-experience-quality
• manage-design-handoff (surface issues to fix before handoff)

Instructions

1. Define the tasks participants will attempt (user-goal-based, not UI-instruction-based).
2. Define success criteria for each task.
3. Select testing method: moderated or unmoderated, remote or in-person.
4. Specify participant criteria and number (5–8 for qualitative usability testing).
5. Design the test script and facilitation guide.
6. After testing: synthesize findings by severity (critical / major / minor).
7. Map findings to specific interface elements.

Output

Provide:

• Test tasks and success criteria
• Testing method and participant criteria
• Test script outline
• (Post-test) Findings organized by severity
• (Post-test) Specific friction points with supporting observations
• Recommended fixes

Risks / caveats

• Tasks must be goal-based, not UI-instruction-based ("book a trip" not "click the Book button")
• Do not coach participants — observe and note, don't guide
• 5 participants typically surface 85% of usability issues; plan accordingly