User interviews: how to plan and run them well
To run good user interviews, set clear research goals, recruit participants who match your users, write a flexible discussion guide, ask open questions about real past experiences, listen more than you talk, avoid leading questions, record with consent and take structured notes. Five to eight interviews per user group usually reveal the main themes.
Why user interviews matter
User interviews are one-to-one conversations with people who use, or might use, a product. They reveal how people think about a problem, what they are trying to achieve, what frustrates them and how they currently cope. Unlike analytics, which show what people do, interviews help explain why. They are one of the most flexible and widely used research methods, valuable at every stage from early discovery to improving a live product.
This guide covers planning and running interviews. A companion post, analysing user interview findings, explains how to turn interview notes into insight and action.
Set clear goals
Every interview study should start with a clear purpose. What decisions will the research inform? What do you need to learn? Write three to five research questions, such as "How do small business owners currently track invoices?" or "What makes first-time buyers hesitate at checkout?" These guide who you recruit and what you ask, and they help keep sessions focused.
Recruit the right participants
Interview people who represent your actual or intended users. Use a short screener to check their experience, behaviour and demographics. For most studies, five to eight interviews per user group reveal the main themes; more may be needed for very diverse audiences. Offer a fair incentive and flexible scheduling. For Gulf audiences, include Arabic and English speakers and different nationalities where relevant. Our guide to UX research agencies in the UAE covers recruitment support.
Write a discussion guide
| Section | Purpose | Example |
|---|---|---|
| Introduction | Explain the purpose and get consent | "There are no right or wrong answers." |
| Warm-up | Build rapport and context | "Tell me a little about your role." |
| Main topics | Explore research questions | "Walk me through the last time you..." |
| Specific probes | Dig into details | "What happened next?" |
| Wrap-up | Final thoughts and thanks | "Is there anything we did not cover?" |
A discussion guide is a flexible outline, not a rigid script. It ensures important topics are covered while allowing the conversation to follow interesting directions. Pilot it with a colleague to check timing and clarity.
Ask good questions
- Ask about the past: "Tell me about the last time you..." rather than "Would you..."
- Keep questions open: Questions that cannot be answered with yes or no.
- Ask one thing at a time: Avoid double-barrelled questions.
- Follow up: "Why was that?" "Can you show me?"
- Avoid leading: Do not suggest the answer you hope for.
- Allow silence: Pauses give people time to think.
People are not good at predicting their own future behaviour, but they can describe what they actually did. Grounding questions in specific past experiences produces more reliable insight. Our post on cognitive biases in user research explains why leading questions distort results.
Moderating the session
- Start by putting the participant at ease
- Listen more than you talk, aiming for the participant to speak most of the time
- Stay neutral; avoid reacting to praise or criticism
- Ask for examples and stories
- Watch for emotions and hesitations
- Keep to time and respect the participant’s schedule
Common interviewing mistakes
- Asking people to design the solution for you
- Pitching the product instead of listening
- Asking leading or yes-or-no questions
- Interviewing only friendly or existing customers
- Skipping notes and relying on memory
Each mistake reduces the honesty or usefulness of what you learn. Reviewing recordings of your own interviews is one of the fastest ways to improve as an interviewer.
Remote interviews
Many interviews now take place over video calls, which makes it easier to reach participants in different locations and lets people join from their own homes or workplaces. Test the technology beforehand, send clear joining instructions, have a backup plan for connection problems and ask participants to share their screen when discussing tools they use. Remote sessions work well for most topics, though in-person interviews can be better for observing physical environments.
For interviews in Arabic, use a native-speaking moderator or a skilled interpreter, and allow extra time for translation and cultural nuance when analysing notes.
Recording and note-taking
With consent, record sessions so you can review them later and share short clips. A second team member can take notes, freeing the interviewer to focus on the conversation. Use a consistent note template, capturing quotes, observations and questions to follow up. Write a short summary immediately after each interview, while details are fresh.
Ethics and consent
- Explain the purpose: Tell participants why the research is happening.
- Obtain consent: Agree to recording and how data will be used.
- Protect privacy: Anonymise data and limit access to recordings.
- Allow withdrawal: Participants can stop at any time.
- Pay fairly: Provide the agreed incentive promptly.
Frequently asked questions
How many user interviews are enough?
Five to eight per user group usually reveal the main themes; more may be needed for diverse audiences.
How long should a user interview be?
Typically 30 to 60 minutes, depending on the topic and participant availability.
What is a discussion guide?
A flexible outline of topics and questions that keeps interviews focused while allowing natural conversation.
Should we record interviews?
Yes, with consent. Recordings help analysis and let you share short clips with stakeholders.
Should stakeholders watch interviews?
Yes, as silent observers. Seeing real users firsthand builds empathy and makes findings more persuasive.
Can we interview users in Arabic and English in the same study?
Yes. Use moderators fluent in each language and analyse both sets of interviews together to compare themes.
How far in advance should we recruit?
Usually one to two weeks, depending on how specialised your participants are.