How to Write a Qualitative Interview Guide, With Example Questions

A qualitative interview guide is the plan for a research conversation. It lists what you need to learn, the order you will ask about it, and the probes you will use when an answer is thin, so every participant gets a comparable interview while the moderator still follows what each person says.

Researchers also call it a discussion guide or a topic guide. The names differ in emphasis, but the job is the same. A good guide reads like a sequence of decisions you need to make, turned into questions a participant can answer from their own experience.

A guide can fail before fieldwork starts if it lists questions without saying what each is for, asks about the brand before the category, or runs long enough that the last section gets rushed. Below is how to avoid that, with example questions and lessons from Listen studies.

What a discussion guide is for

A discussion guide does two jobs at once. It keeps interviews consistent enough that you can compare answers across participants, and it gives the moderator room to follow an unexpected answer. Too loose a guide produces interviews you cannot compare, and too tight a guide turns into a spoken survey.

The guide carries even more weight when an AI interviewer runs the sessions, because the guide is the moderator's whole briefing. Our study comparing AI interviewers and human moderators concluded that in a human-moderated interview the risk is self-censorship, and in an AI-moderated one the risk is a shallow follow-up. "Only the second can be fixed with a better discussion guide." In the same study, 8% of participants said humans are better at probing and clarifying, the gap that well-written probes close.

Start with research objectives

Write three to five objectives before you write a single question. Each objective should name a decision someone will make with the answer, such as which positioning to lead with or which message to cut. Then map every question to an objective. A question that serves none of them goes.

Objectives also decide what the guide measures and what it explains. Our apparel brand positioning research had two stages, territory importance and message testing, so the guide had participants rate four territories, then react to 12 messages and explain each reaction.

How to structure a qualitative interview guide

Interview guides for qualitative research usually move from easy and general questions to specific and evaluative ones:

1. Introduction. Purpose, length, no wrong answers, consent to record.

2. Warm-up. Easy questions about the participant's routine in the category, to get them talking and give you context.

3. Category behavior. How they choose, what frustrates them, what the last purchase looked like.

4. Core topics. The questions tied most directly to your objectives.

5. Stimulus. Concepts, messages, ads or prototypes, shown after the unprompted questions.

6. Wrap-up. What they would change and anything you did not ask about.

Order matters because early questions shape later answers. The apparel study's advice was to "start with the category, not the brand," so participants describe benefits in their own words before seeing yours. Our Super Bowl ad testing study moved from memory to behavior: unprompted recall first, then brand opinion, then action, and only at the end did viewers watch four ads on camera. Showing the ads first would have contaminated the recall questions.

Writing open questions and probes

Good open questions ask about a specific experience. "Tell me about the last time you switched brands" gets you a story with details. "Why do people switch brands" gets you a theory. Ask one thing per question, avoid words that suggest the answer you want, and save "why" for after the participant has described what happened.

Write probes into the guide instead of leaving them to the moderator's instinct. Common probes include asking for an example, asking what happened next, asking what they compared it with, and repeating a word they used back to them. Our restaurant message testing study paired every rating with a follow-up. Diners rated each message, then explained in their own words what worked or what didn't, and the study's advice was to "ask why after every rating."

If you are worried about steering answers, see our guide to types of bias in research, which covers leading questions and moderator effects in more detail.

Open-ended qualitative interview questions examples

These work across most consumer topics. Adjust the category and time frame to your study.

• Walk me through the last time you bought something in this category. Where were you, and what did you look at first?

• What made you pick that one over the others you considered?

• Tell me about a time a product in this category disappointed you. What happened?

• If you could design the ideal brand here, what would it do differently?

• Which brands come to mind when you think of this category? What do you know about each one?

• After seeing this message, what is it telling you? What, if anything, do you not believe?

• What would you do next after seeing this, if anything?

• Is there anything I should have asked about but didn't?

Stimulus, timing and interview length

Plan time per section, then cut. Guides almost always run long on paper and longer in the field. Put the questions tied to your most important objective early, because attention drops and some participants will not finish. In the Super Bowl study, 56 viewers answered the recall questions and 51 the celebrity and AI questions, so each percentage used its own base. Write the guide expecting that drop-off.

Show stimulus one piece at a time and rotate the order so no item always benefits from going first. In the restaurant study each diner had a 15 to 20 minute AI-moderated video interview and saw all 12 messages, one category at a time, which kept comparisons clean.

Using AI to write a discussion guide

Researchers are skeptical. In our study of 76 market researchers, 91% were open to AI in their workflow, but only 18% trusted AI-written discussion guides, surveys and stimuli as much as traditional research. Of those who described what AI would need to earn their trust, 65% said it must prove it's accurate and 52% wanted a human checking the output.

Use AI to draft a first version, suggest probes and check coverage against your objectives. Keep the objectives, question order and final wording with the researcher. The same rule applies to analysis, as our guide to qualitative data analysis explains.

On Listen, the researcher writes the guide and an AI interviewer runs it with every participant, asking the follow-ups you specify and probing vague answers. That lets you run in-depth interviews at survey scale while keeping the guide in your hands.

FAQ

What is a qualitative interview guide?

A qualitative interview guide is a document that sets out the objectives, question order and probes for a research interview. It keeps interviews comparable across participants while leaving the moderator free to follow up on what each person says.

What is the difference between a discussion guide and a topic guide?

In practice the terms are used interchangeably. Some researchers use "topic guide" for a looser list of themes to cover and "discussion guide" for a fuller document with worded questions and probes.

How many questions should an interview guide have?

Fewer than you think, because follow-ups take more time than the main questions. Budget minutes for each section and drop any question that doesn't map to an objective.

What are good open-ended qualitative interview questions?

Good questions ask about a specific recent experience, such as "Walk me through the last time you bought this." They ask one thing at a time, avoid suggesting an answer, and come with written probes like "Can you give me an example?"

Can AI write a discussion guide?

AI can draft a guide and suggest probes, but researchers remain cautious. In Listen's study of market researchers, only 18% trusted AI-written discussion guides, surveys and stimuli as much as traditional research, so a researcher should own the objectives and final wording.