Usability Testing

What is it?

Usability testing is a research method in which representative users are asked to complete realistic tasks using a product (or prototype) while researchers observe, listen, and measure. It identifies where users struggle, why they struggle, and what they do instead of what designers expected. Unlike surveys or analytics, usability testing reveals the reasoning behind behaviour — the "why" behind the "what".

Why it matters

Analytics show you where users drop off. Usability testing shows you why. A 5-user usability test typically uncovers 85% of major usability problems (Nielsen, 1993). It is one of the highest-ROI research methods available: catching a single critical usability failure before launch can save hundreds of hours of post-launch fixes, loss of users, and brand damage. It is also the only method that puts actual human behaviour — not assumptions — at the centre of design decisions.

Best Practices

• Test with 5 users per round to find ~85% of usability problems. Run multiple small rounds rather than one large one.
• Write task scenarios using realistic user goals — "Find and book a hotel for 2 nights in Lisbon in July" — not instructions — never say "click the search button".
• Use the think-aloud protocol: ask participants to narrate their thoughts as they work. This reveals mental models, confusion, and expectations.
• Recruit participants who match your real user demographics and technical fluency. Testing with developers is not the same as testing with your actual users.
• Observe without intervening. Resist the urge to help when participants struggle — that struggle IS the data.
• Record sessions (with consent). You cannot watch, note, and analyse simultaneously. Video evidence also convinces stakeholders.
• Measure both qualitative data (verbal confusion, misunderstandings) and quantitative data (task completion rate, time-on-task, error count).
• Run unmoderated testing (e.g., Maze, UserTesting.com) for speed and scale. Run moderated testing for depth and nuance.
• Test prototypes early — even paper prototypes reveal navigation and flow problems before a single line of code is written.
• Debrief immediately after each session. Memory of nuanced moments fades within hours.

Common Mistakes

• Writing leading tasks: "Use the search feature to find a hotel" tells users what to do instead of giving them a goal.
• Testing with colleagues or teammates who know the product — familiarity blinds them to the friction real users experience.
• Running one large test at the end of a project instead of iterative small tests throughout.
• Focusing only on whether tasks are completed, not on how — the path reveals mental model mismatches.
• Treating usability testing as a checkbox activity rather than acting on the findings.
• Recruiting too few or too many participants — 5 is enough for qualitative; 20+ is needed for statistical significance in quantitative studies.
• Asking participants if they "like" the design — preference is not usability.

Checklist

Research & Theory

Real-World Examples