Usability Testing interview questions
Usability Testing questions a UX / UI designer can be asked, from first job to senior. Open one to see the strongest answer and why it works.
A slide says “92% usability.” You ran no study. What do you do?
The strongest answer: Take the number off — say what we watched, I do not keep a costume 92
Why it works: Exactly. No study, no 92. Decimals and sourceless SUS are worse costumes.
- Remove invented usability %.
- Say what you watched.
- Decimals ≠ a study.
- No sourceless industry SUS.
What is the difference between a moderated and an unmoderated usability test, and when do you pick each?
The strongest answer: Moderated means I am there and can ask why, which suits a new or messy flow; unmoderated is faster at volume and suits a short task I can write without any ambiguity
Why it works: Right — the trade is depth and follow-up questions against speed and numbers.
- Moderated buys “why”.
- Unmoderated buys volume.
- Messy flow: sit with them.
- Clear task: send it out.
What does a usability test need besides a prototype?
The strongest answer: A task in their words, a think-aloud, and notes on where they stall — I do not pitch the design while they try
Why it works: Exactly. Their task, their stalls. Pitches, NPS-in-the-lab, and pre-written success rates are theatre.
- Task in their words + stalls.
- Do not pitch during the attempt.
- Do not swap in an NPS.
- No pre-written success %.
Five of five stalled on the same step. What do you write?
The strongest answer: “Five stalled here” and what we will change — I do not write “73% of users fail”
Why it works: Exactly. Count what you saw. 73%, ignore-because-n, and invented SUS are three lies.
- Write the count you saw.
- Do not invent 73%.
- Do not ignore a shared stall.
- No invented SUS.
You need to test a payroll flow. Who do you not book?
The strongest answer: Random friends who have never run payroll — I book people who do the job, even if fewer
Why it works: Exactly. The job. First-replies, builders, and yourself are how you test a hallway.
- Recruit people who do the job.
- Fewer right > many wrong.
- Do not only test builders.
- Do not only test yourself.
What do you standardise in usability across teams?
The strongest answer: Consent, task style, and how findings reach a decision — I do not standardise every prototype tool
Why it works: Exactly. Consent and a path to a decision. One-sitter, novels, and invent-every-time miss the bar.
- Standardise consent, task style, decision-path.
- Let tools differ.
- Do not require one sitter on every test.
- Do not write a shelf protocol.
Your team has run tests every sprint for six months; the report folder is full and the product is unchanged.
What do you change?
The strongest answer: Turn each finding into a named change with an owner and a re-test date, prioritise the few that block the task, and stop starting new studies until the last one has moved something
Why it works: Exactly — a finding with no owner and no re-test is a document, and closing the loop is the whole point.
- Close the loop, not the report.
- One finding, one owner, one change.
- Prioritise what blocks the task.
- Re-test the change.
How do you talk about a redesign’s effect without a fake lift?
The strongest answer: The task we watched, what changed, and what we did not measure — I do not print +34% UX
Why it works: Exactly. Task, change, leftover. +34%, cab-SUS, and heatmap-joy are costumes.
- Name the task and the change.
- Say what you did not measure.
- No +34% UX.
- Heatmaps ≠ joy.
Who owns the usability practice?
The strongest answer: A named owner, a path into decisions, and a budget to pay participants — not “designers will test when they can”
Why it works: Exactly. Operated. First-author, vendor crowns, and marketing-as-owner miss the gate.
- Name an owner and a decision path.
- Fund paying participants.
- Vendor crowns are not your practice.
- Do not give it to marketing.
The board wants one number for design quality. What do you give?
The strongest answer: A task we watch on a real flow, plus a leftover — I do not invent a design-health %
Why it works: Exactly. A watched task and a leftover. Awards, counts, and crowns are games.
- A watched task on a real flow.
- One leftover sentence.
- No invented health %.
- No award as the answer.
What leftover do you put on a go/no-go for a money flow?
The strongest answer: Which states we have not tested with a real person, and who accepted that — I do not say “it feels solid”
Why it works: Exactly. Untested states and an accepter. Premium, counts, and invented SUS are not a gate.
- Untested states + who accepted.
- Not “feels solid.”
- Not a frame count.
- No invented SUS.