During a usability session, a participant stops at a form and says, “I wouldn’t put that in.” A recording captures the hesitation, but a researcher who is present can discover whether the objection concerns privacy, relevance or simply not having the information to hand. That opportunity to follow an unexpected moment is the main reason to choose moderated testing.
In a moderated test, a researcher accompanies a participant as they attempt tasks with a product or prototype. The session may happen in person or remotely. The researcher introduces the activity, observes the attempt and asks questions where they help explain what happened, while taking care not to teach the participant how the design works.
When being there changes what you can learn
Moderation is particularly useful when a task has several possible routes, the prototype is incomplete, or the team needs to understand a decision that clicks alone cannot explain. A researcher can also adapt the practical arrangements to a participant’s access needs, although a live call is not automatically the most accessible or comfortable format for everyone.
Consider a fictional expense tool being tested with people who submit claims for work. A participant leaves a receipt unsubmitted even though the interface seems straightforward. A follow-up reveals that they are waiting for a manager to confirm the cost code. The obstacle belongs partly to the organisation’s process, and a different button label would leave it unresolved. That context may warrant further user interviews beyond the usability session.
Moderation carries its own influence. A participant can look to the researcher for reassurance, infer approval from their tone or persist longer than they would alone. GOV.UK’s advice for moderated usability testing offers practical guidance on keeping the session focused on the participant’s experience.
Prepare the session, including the moments of difficulty
Write the research questions and task scenarios before drafting a long list of prompts. Recruit people whose circumstances matter to those questions, prepare realistic test information, and rehearse the route through the prototype. Participants should know what is being recorded, who will observe and how their contribution will be used.
Agree when assistance will be offered. If the study measures independent completion, help changes the meaning of the result and must be recorded. If the purpose is to examine a later stage of the journey, it may be reasonable to note an early failure and move the participant forward. Neither choice should depend on how uncomfortable the moderator happens to feel during a particular pause.
Allow room in the schedule for consent, technical checks and a closing conversation. A session packed with tasks can leave no time to understand the very incident that makes the research valuable. For remote studies, test the screen-sharing arrangement on the device participants will actually use; live interviews in Ballpark are one option for bringing the researcher and participant together.
Ask about the experience without supplying it
“What are you looking for?” leaves room for the participant’s account. “Are you looking for the receipts tab?” supplies an interpretation and a route. The distinction is easy to recognise in a script and harder to maintain when someone appears stuck in front of you.
Use silence deliberately, and distinguish a request for clarification about the research from a request for help with the product. Explaining what a task means may be necessary; explaining which control fulfils it changes the task. If you use a think-aloud protocol, keep reminders neutral and avoid making the session a continuous interview while the person is trying to work.
Keep the moderator in the record
Analysis should include what the researcher said and did, particularly around hesitation, errors and assistance. A task completed after a hint is evidence about assisted use. It should not quietly join the tally of independent successes.
The same care applies when several researchers share a study. Compare their introductions, prompts and intervention rules before treating the sessions as equivalent. Moderated testing offers a richer conversation around observed behaviour, but the credibility of the findings depends on remembering that the researcher was part of the situation being observed.