An interview guide is most useful at the moment a conversation becomes interesting. When a participant raises an unexpected difficulty, a prepared researcher can follow it and still know which topics remain uncovered. Without that preparation, it is easy to spend the session pursuing whichever detail arrived last, or to retreat into a list of questions that no longer fits the conversation.
The guide is a working document containing the study’s objectives, topics, opening questions, and possible follow-ups, together with the practical arrangements for the session. In semi-structured interviews, it supports a conversation rather than prescribing every word.
Turn a research objective into questions people can answer
A research question belongs to the team; an interview question needs to make sense to the participant. “What prevents adoption?” may be a useful starting point for the business, but it asks a customer to supply an analysis. A prompt about the last time they tried to use the product gives them something more concrete to describe.
Suppose a fictional team wants to understand why people stop preparing a report halfway through. The guide could begin with the circumstances around the latest report, then explore the information required, the steps attempted, interruptions, and what happened afterwards. These topics help investigate abandonment without assuming that the interface was the cause.
GOV.UK’s guidance on planning a research round connects research objectives with session topics. Use that connection to decide what belongs in the guide: if a question cannot contribute to the study’s purpose, reconsider whether it needs time in this session.
A short example of a working guide
For that reporting study, the opening would explain the purpose of the session, confirm the participant’s agreement to the proposed recording, and establish their role. An initial question might be: “Tell me about the last report you needed to prepare for someone else.”
Under “Getting started”, possible follow-ups could ask who needed the report, what information was required, and where the participant looked first. Under “Completing the work”, the researcher could ask about anything they had to check, anything that interrupted progress, and how they decided the report was ready. A final topic would cover what happened after it was sent, including any corrections or further requests.
The guide should distinguish these optional probes from the essential topics. A participant who has already explained an interruption does not need to be asked the same question again merely to complete the document. Keep any assessment criteria or private research notes separate from the wording addressed to the participant.
Wording and order affect the conversation
Questions are easier to answer when they contain one clear idea and do not assume an experience the person may not have had. “What frustrated you about exporting?” presumes frustration, while “How did exporting go?” leaves more room for an uneventful experience. Specific follow-ups can then explore whatever the participant actually describes.
Order matters too. Showing a proposed feature before discussing existing work can influence the terms people use and the problems they emphasise. If both are relevant, establish a clear transition between exploring current experience and evaluating the proposal, so that the analysis can distinguish the two.
Pilot the guide and record revisions
Read the questions aloud, then try the session with someone suited to the study. Check whether prompts are understood, whether the topic order feels coherent, and whether enough time remains for the most important material. A pilot is also an opportunity to discover that the research question needs narrowing.
Keep a dated record of substantive changes and note which sessions used each version. If a new prompt introduces an issue halfway through the study, the later answers cannot be interpreted as though everyone had the same opportunity to discuss it. This record becomes part of the evidence when analysing the user interviews, helping the team explain both what participants said and how those accounts were obtained.