Interview Guide for UX Research

When to use it

For semi-structured interviews and for usability tests with a guided walkthrough. It is the document you take into the session: what to ask, in what order, and how much time each block gets.

It is not for surveys or for entirely unstructured conversations. If your conversation has no blocks and no timings, it is not a guide, it is a list of topics.

It takes two to four hours to build, depending on how clear your objective is. Most of that goes into the step people skip: writing the research questions before the interview questions.

Before you use it

Research questions and interview questions are not the same thing. The first are what you need to find out — three to five, written first. The second are what you actually say to the person, and they rarely resemble each other. "Which step do they abandon the signup at?" is a research question; what you ask the person is "tell me what happened the last time you tried to create an account".

The timings are an allocation, not a stopwatch. They exist so you know you have overrun, not so you cut someone off mid-way through something good.

This guide does not replace consent. The opening block includes the script you read aloud, but the written authorization is a separate document: the informed consent template is right next door.

Everything in brackets gets replaced. The worked example lives on this page and not in the file you download, so there is nothing to remember to delete.

The document

This is what you download and take to the session.

Study: [study name] Research question: [what you need to find out, in one sentence] Moderator: [name] · Observer: [name] Length: [60] minutes · [remote over video call / in person at (place)] Field dates: [DD/MM/YYYY – DD/MM/YYYY]

1. Project details

2. How to use this guide

• Replace everything [in brackets]. If a bracket is still there when you review it, the guide is not ready.
• Blocks labelled optional get deleted whole if they do not apply. Do not leave the heading orphaned.
• Use the document's real headings, not simulated bold: that is what builds the table of contents.

Splitting rule: sections 1, 2, 3, 11 and 12 are for the team only. Sections 4 to 10 are the script you read out loud, and they are the only ones the participant version contains.

Before you start

• Signed consent received. Without it, the session is not recorded and does not enter the analysis.
• Recording running.
• [Link to the prototype / test environment] open and working.

3. Session objective

[What you need to understand by the end, in two or three sentences. This is not the project objective: it is this conversation's.]

Out of scope: [what this session will not answer, said now]

4. Opening and consent · 5 min

Read out loud, with the recording already running.

"Hi [name], thank you for your time. I'm [name] and I work in user research for [organization]. We're going to talk for about [60] minutes about [topic].

There are no right answers. We're not evaluating you, we're evaluating [the product]. If something seems confusing or badly done, saying so is exactly what we need.

I didn't design this, so you won't offend me.

You can skip any question or ask me to stop, without explaining why. Any questions before we start?"

5. Warm-up and context · 5 min

Easy questions, so the person hears themselves talk before what they say matters.

1. Tell me what you do and what a normal week looks like for you.
2. [Question about the context the product lives in, without naming it yet.]
3. [Question about how they solve the problem today, with whatever they use.]

6. Block 1 — [block name] · [10] min

One block per research question. Three at most: with more, none of them gets deep enough.

1. [Open question: "tell me about the last time you…"]
2. [Open question]
3. [Open question]

7. Block 2 — [block name] · [10] min

1. [Open question]
2. [Open question]
3. [Open question]

8. Block 3 — [block name] · [10] min

1. [Open question]
2. [Open question]
3. [Open question]

9. Optional block — Guided walkthrough · [20] min

Delete it whole if the session does not include a walkthrough. Instruction read as written:

"Now I'm going to ask you to use [product] and move through it as if you were on your own. Talk me through what you're seeing and what you expect to happen. If you get stuck, don't ask me for help: stay there and tell me what you would do."

Completion criterion for each task: [what counts as done]. If you are going to measure time or success per task, that gets recorded in the usability test protocol, not here.

10. Closing · 5 min

1. If you could change one single thing about all this, what would it be?
2. Is there anything I didn't ask that you think I should know?

"Thank you very much. [If there is an incentive: your incentive of (amount) arrives via (channel) within (timeframe).] If you remember something afterwards, write to me at [contact]."

11. After the session

Filled in within the hour, while the conversation is fresh.

12. Method note and limitations

[What this study will be able to say and what it will not. If it is qualitative: it describes reasons and mechanisms, not proportions.]

It expires. Review on [date].

Probes

To keep at hand and use in any block.

• Can you give me an example?
• What happened next?
• What made you think that?
• Tell me more about that part.
• How did you work around it?
• You said [word they used]. What do you mean by that?

The worked version

The same document with a case written end to end: an invented study on the sign-up flow of a personal finance app. It shows what level of detail each field expects before you face the empty version. It downloads separately.

Internal version. This file contains method notes and team decisions. It is not shared with participants. What the participant sees is a separate file: Interview guide (participant).

1. Project details

2. How to use this guide

• Replace everything [in brackets]. If a bracket is still there when you review it, the guide is not ready.
• Blocks marked EJ are examples written against the reference case. Delete them whole before taking the guide into the field.
• Blocks labelled OPTIONAL get deleted whole if they do not apply. Do not leave the heading orphaned.
• The automatic table of contents in Google Docs is built from this document's real H1/H2/H3 headings. If you add a section, use the heading style, not bold text.

Splitting rule: sections 2, 3, 9 and 10 are for the team only. Sections 4 to 8 are the script you read out loud. The participant version contains only 4 to 8.

3. Session objective

Understand what happens between the moment someone downloads the app and the moment they decide to abandon sign-up. We care about the real order of the steps, what information the app asks for that the person does not have at hand, and at what point distrust appears.

Out of scope: [advanced use of the app, investment features, customer support].

4. Opening and consent

Read out loud

Time: 5 min

Hi [name], thank you for your time. I'm [name] and I work in user research for [client]. We're going to talk for about [60] minutes about how you use apps to manage your money.

There are no right answers. We're not evaluating you, we're evaluating the app. If something seems confusing or badly done, saying so is exactly what we need.

I'm going to record audio and screen so we don't lose detail. We use the recording inside the working team only, it is not published and it is not tied to your name. You can ask me to stop at any moment, or skip a question without explaining why. Are you okay with us recording?

Wait for explicit verbal confirmation before starting the recording.

Signed consent: [link to the form]. With no consent on record, the session is not recorded and does not enter the analysis.

5. Warm-up and context

Time: 8 min

1. Tell me what you do and what a normal week looks like for you.
2. How do you keep track of your spending today? Show me, if you have something at hand.
3. What banking or finance apps do you have installed on your phone?
4. Of those, which one do you open most often? What do you open it for?

EJExample of a useful follow-up — delete this block

If they answer "I write everything down on my phone": Can you show me where you write it down? When was the last time you did? Ask for the concrete evidence instead of the general description.

6. Block 1 — Before downloading

Time: 10 min

1. How did you hear about [app name]?
2. What did you expect it to do for you before you installed it?
3. What made you decide to download it that particular day?
4. Before opening it the first time, what did you think it would ask you for?

Probes

• What made you think that?
• Had you tried something similar before? How did that end?
• Tell me more about that part.

7. Block 2 — The activation journey

Time: 20 min · with the app on screen

Instruction to the participant: "Now I'm going to ask you to open the app and move through it as if it were the first time. Talk me through what you're seeing and what you expect to happen. If you get stuck, don't ask me for help: stay there and tell me what you would do."

1. Before you touch anything: what do you think this screen is, and what is it asking you for?
2. [They move on] What did you expect to happen when you pressed that button? Did that happen?
3. At this step it asks you for [requested detail]. Do you have it at hand right now?
4. If you were on your own at home, what would you do at this point?
5. Is there anything on this screen you don't understand or that makes you hesitate?

EJExample row, filled in — delete this block

Step 5 · Bank linking — "Oh no, not this. It's asking me for my bank password inside another app. This is where I leave." Closes the app without completing. Opens it again twice and stops at the same point.

8. Block 3 — The moment of abandonment

Time: 12 min

1. Let's go back to the point where you stopped. What were you thinking there?
2. What would have had to happen for you to keep going?
3. After closing the app, what did you do? Did you try again?
4. If a friend asks you whether it's worth installing, what do you tell them?
5. If you could change one single thing about the whole process, what would it be?

Do not offer solutions or defend the design. If they ask for your opinion, hand the question back: "what would you do?".

OPTIONAL BLOCK — DELETE IT WHOLE IF IT DOES NOT APPLY

9. Optional block — Trust and personal data

Time: 8 min · use only if the session is running ahead of schedule

1. What details do you think it's reasonable for a finance app to ask you for?
2. And what would you never give it? Why that one?
3. How do you tell whether an app is trustworthy or not?

10. Closing

Time: 5 min

Before we finish: is there anything I didn't ask that you think I should know?

Thank you very much. Your incentive [amount/type] arrives via [channel] within [timeframe]. If you remember something afterwards, write to me at [contact].

11. After the session

Team only

• Write the debrief while it is hot, before the next session: three things that surprised you and one verbatim quote.
• Mark the step where they stopped in the findings and severity matrix.
• Record whether you had to skip questions and why. That affects comparability across sessions.

12. Method note and limitations

Semi-structured interviews with a guided walkthrough, [n] participants, [remote over video call]. The order of the blocks stays the same; the exact wording of each question adapts to the conversation.

Limitations: the walkthrough is retrospective and reconstructed, not observation of real first use. The findings describe reasons and expectations, not frequencies: they are not used to estimate drop-off rates. [n] sessions do not allow segmenting by [age / income level].

UXR SpA · User research [contacto@uxr.cl] · [uxr.cl]

Template v1.0 · reviewed [DD/MM/YYYY] Original at UXR — check that your copy has not expired

The participant version

What the other person receives: the conversation script or the form, with no method notes, no screening criteria and none of the team's hypotheses. It is written over the same example case, so its main use is seeing where the line falls between what gets shared and what does not. It downloads separately.

Study: [study name] · Date: [DD/MM/YYYY] Length: [60] minutes

Version to share. It contains only the conversation script: sections 4 to 10 of the internal guide. No team objectives, no method notes, no observation criteria.

What we are going to do

[In two sentences and without jargon: what the conversation is about and what the person will be asked to do.]

• There are no right answers. We are evaluating [the product], not you.
• You can skip any question or stop the session at any point.
• [If it is recorded: we record (audio / audio and screen) only so we don't lose detail.]

1. Opening and consent · 5 min

"Hi [name], thank you for your time. I'm [name] and I work in user research for [organization]. We're going to talk for about [60] minutes about [topic].

There are no right answers. We're not evaluating you, we're evaluating [the product]. If something seems confusing or badly done, saying so is exactly what we need.

I didn't design this, so you won't offend me.

You can skip any question or ask me to stop, without explaining why. Any questions before we start?"

2. To get started · 5 min

1. Tell me what you do and what a normal week looks like for you.
2. [Question about the context, without naming the product yet.]
3. [Question about how they solve the problem today.]

3. [Name of block 1] · [10] min

1. [Open question]
2. [Open question]
3. [Open question]

4. [Name of block 2] · [10] min

1. [Open question]
2. [Open question]
3. [Open question]

5. [Name of block 3] · [10] min

1. [Open question]
2. [Open question]
3. [Open question]

6. [Optional block] · [20] min

Delete it whole if it does not apply.

"Now I'm going to ask you to use [product] and move through it as if you were on your own. Talk me through what you're seeing and what you expect to happen. If you get stuck, don't ask me for help: stay there and tell me what you would do."

• [Task 1]
• [Task 2]
• [Task 3]

7. Closing · 5 min

1. If you could change one single thing about all this, what would it be?
2. Is there anything I didn't ask that you think I should know?

"Thank you very much. [If there is an incentive: your incentive of (amount) arrives via (channel) within (timeframe).] If you remember something afterwards, write to me at [contact]."

How to moderate with it

The guide is an anchor, not a script. It exists so every session covers the same ground and nothing gets forgotten, not to be read straight through. If the person goes somewhere interesting, let them; the guide is there to come back to.

Open questions, not closed ones. "Did you like it?" gets one word and leaves nothing. "Tell me what happened" opens. Quick test: if your question can be answered yes or no, rewrite it.

The silence is yours, not theirs. After a short answer, waiting three seconds usually brings the part that matters. It is uncomfortable and it works.

Do not defend the design. If they ask what it should do, hand the question back: "what would you do?".

Pilot it. One or two practice sessions before the first real one. That is where you find out you had twice as many questions as fit, that two of them ask the same thing, and that the third one does not parse when said out loud.

A worked example

A study on the onboarding of a personal finance app. The research question is what stops people from activating their account for the first time.

The blocks came out as: before downloading (how they heard about it, what they expected), the activation walkthrough (with the app on screen, moving through it as if for the first time) and the moment they stopped (what they were thinking, what would have had to happen to continue).

One question from that study, to see the difference between the two layers: the research question was "what information does the app ask for that the person does not have to hand?". What the person was asked, at the step where it happened, was: "at this step they ask you for [data]. Do you have it to hand right now?".

This example lives here only, not in the downloadable file.

Check before using it

• Did you write the research questions before the interview questions?
• Can any question be answered yes or no?
• Are there more than three blocks in the main section?
• Do the timings add up to the session length, or to more?
• Any bracket left unreplaced?
• Did you pilot it with someone?

Grounding

The structure — opening with a statement of impartiality, warm-up, main discussion or tasks, closing — and the quality criteria come from this site's Interview Guide entry, where they appear in more detail and with their sources.

The opening block covers what article 3 of Chile's Law 19.628 requires: informing people of the purpose and of whether answering is optional. The written authorization is a separate document: the informed consent template.