A grand tour question asks a participant to walk you through a whole piece of their world: a workday, a recurring process, a place, or a sequence of events.
For example:
Walk me through how month-end normally works, from the first thing you open to the point when the numbers are final.
That question gives you a map. Then you make the map concrete: ask about the last month-end, zoom into one or two steps with mini-tour questions, and—when it is safe and the participant has agreed—watch part of the work over a screen share.
This sequence matters because your product is rarely the whole workflow. It may be one tab between a spreadsheet, an email thread, a colleague’s approval, and a workaround nobody thought to mention.
What is a grand tour question?
Anthropologist James Spradley described grand tour questions as descriptive prompts that take the interviewer through a setting, time period, sequence, group of people, activity, or set of objects. The participant knows the territory; the interviewer asks to be shown around in the participant’s terms.
The technique transfers neatly to B2B product research:
- “Walk me through a normal handover from sales to implementation.”
- “Take me through the last time you closed the books.”
- “Show me how a support request moves from the inbox to resolution.”
- “Sketch the tools you use to prepare a board report and show me where information moves between them.”
Beth Leech called the grand tour the most useful question she knew for a semistructured interview because it can produce a broad answer and reveal which follow-ups matter. Treat that as experienced practitioner guidance, not a guarantee that every participant will give you a long or useful response. Some people need a narrower starting point.
The practical value is simple: you learn the participant’s workflow before forcing it into your product’s categories.
Start with the decision, not the product
Before you write the question, decide what the interview should help you choose.
Suppose your research question is:
Where does invoice reconciliation break down, and which part should we improve first?
Do not ask, “What should we add to reconciliation?” or “How do you use our export?” Those prompts begin inside the solution. A grand tour opens the surrounding process:
Walk me through how your team reconciles invoices at month-end.
Teresa Torres distinguishes research questions from interview questions. The research question is what your team needs to learn. The interview question is the prompt that helps a participant recall an experience containing relevant evidence.
Scope the tour tightly enough to serve the decision. “Walk me through your job” may be too broad. “Walk me through how you investigate a failed payment” is narrow enough to fit an interview but wide enough to expose tools, people, and handoffs beyond your product.
This also extends the principle in our guide to user interview questions for SaaS founders: what you want to know is not always what you should ask directly.
The four types of grand tour question
Spradley described four forms. You can use more than one in the same study.
| Tour type | Example | What it gives you | Main limitation |
|---|---|---|---|
| Typical | “Walk me through a normal month-end.” | The overall stages and the participant’s own categories | May produce an idealised or averaged account |
| Specific | “Take me through the last month-end.” | A real sequence with people, tools, and exceptions | Recall is still incomplete |
| Guided | “Could you show me how you do the reconciliation?” | Actions, tool-switching, artefacts, and habitual details | Observation can change behaviour and expose sensitive data |
| Task-related | “Sketch the tools involved and draw where the data moves.” | A concrete prompt that can jog memory and expose relationships | The artefact is the participant’s representation, not a direct observation |
The useful pattern is usually typical → specific → guided or task-related.
Nielsen Norman Group treats a typical grand tour as a reasonable way to set the scene at the start of an interview, then recommends moving to specific examples. The typical tour is a draft map. The specific tour tests that map against one episode.
Do not make the opposite mistake and ask only for a dramatic story. The critical incident technique is useful for important positive or negative events, but people tend to recall extremes. Small irritations and ordinary usage may disappear. A workflow tour helps you hear about the unremarkable Tuesday where most product use happens.
For a fuller explanation of why “usually” is not the same as a remembered event, see how memory shapes user accounts.
Zoom in with mini-tour questions
A grand tour gives you stages. A mini tour takes one stage and asks for the smaller sequence inside it.
Spradley illustrates this with research on directory-assistance work. A call averaged 37 seconds, but when an operator described what happened inside one call, that short period contained more than a dozen activities. The top-level label—“taking calls”—hid the work.
Your participant may do the same:
“Then I export it and send it to finance.”
Treat that summary as a signpost. Ask:
- Sequence: “What is the first thing you do after the export finishes? What happens next?”
- Complete list: “Where can the supporting documents come from? What else?”
- Example: “Can you give me an example from the last close?”
- Handoff: “Who receives it? How do they know it is ready?”
- Artefact: “What does the file or message look like at that point?”
- Exception: “What changes when something is missing?”
Steve Portigal’s interviewing-question types are useful here because they give you several neutral ways to probe without supplying the answer.
Ask one question at a time. “Walk me through it and tell me what was difficult” is a compound prompt: the participant may answer only one half, and you have already framed the process around difficulty. Let the sequence emerge first. Ask about a hard moment later.
Capture the participant’s language
Grand tours do more than reveal steps. They reveal what people call those steps.
Spradley used native-language questions to reduce the amount of translation participants did for the researcher. In product research, that means noticing when a participant says “the checking sheet”, “the exceptions queue”, or “ready for finance” instead of using the labels in your interface.
Useful prompts include:
- “What do you call that with your colleagues?”
- “When would someone say it is ready?”
- “You called it the exceptions queue. What goes into it?”
- “Is there another name people use for that?”
Avoid treating one person’s phrase as a naming decision. Record it verbatim, then look for the same concept and language across relevant participants. Nielsen Norman Group’s match-between-system-and-reality heuristic recommends familiar words and concepts, but your interface still has to work for the whole target audience.
Keep participant wording separate from your interpretation in your notes. The guide to taking notes in a user interview includes a lightweight way to preserve that traceability.
Add observation when the workflow allows it
A spoken tour is reported behaviour. A guided tour lets you observe at least part of the process.
That distinction can be material. In one Nielsen Norman Group contextual-inquiry example, insurance specialists described copying data from a spreadsheet into another system. Observation revealed extra steps: they also checked a second tool for missing information and repeatedly clicked Save even though the software saved automatically.
Government Digital Service found comparable gaps while observing teams using GOV.UK Notify. Staff in some workplaces could not use phone-based authentication because their phones stayed in lockers. Some teams maintained separate spreadsheets of sent messages. The context changed what the product needed to support.
Remote contextual inquiry works best when the activity is computer-based, desk-based, and visible on a shareable screen. It is not the right method for every workflow. Before the session:
- Explain exactly what you hope to observe and obtain informed consent.
- Check whether screen sharing or recording requires employer approval.
- Ask the participant to close unrelated confidential or personal material.
- Agree how to pause, hide, or skip sensitive steps.
- Test the call and screen-sharing setup.
Do not surprise someone by asking for their whole desktop during the call. For customer data, financial records, health information, or internal messages, use a redacted example, a safe test account, or a spoken reconstruction instead.
During the work, ask short questions in context: “You switched tabs—what is that one for?” Kate Kaplan’s remote contextual-inquiry guidance recommends active inquiry when a detail may be hard to reconstruct later, while acknowledging that every interruption changes the natural flow. Use judgement and let the participant keep working when the action is self-explanatory.
Observation strengthens an interview; it does not make it objective or complete. The participant knows they are being watched, and a short session captures only a slice of their work. Our guide to observed versus reported behaviour explains how to combine the two without overstating either.
A 45-minute grand tour interview guide
Use this as a planning template, then pilot it. The timings are allocations, not evidence-based rules.
1. Frame the study — 3 minutes
Explain the purpose, recording or observation terms, and what the participant can skip. Ask about their role only as far as it helps you understand the workflow.
2. Typical grand tour — 7 minutes
Walk me through how [workflow] normally happens, from the point it begins to the point you consider it finished.
Capture stages, tools, people, handoffs, and the participant’s words. Do not interrupt every stage with a mini tour.
3. Specific grand tour — 10 minutes
Now take me through the last time you did it. What started the process?
Continue chronologically. Notice where the real episode differs from the typical account.
4. Two mini tours — 12 minutes
Choose the stages most relevant to the decision or most surprising in the specific account. Ask for sequence, artefacts, handoffs, exceptions, and workarounds.
5. Guided or task-related tour — 10 minutes
If it is safe and agreed, observe one short part of the workflow. Otherwise, ask the participant to sketch the tools and information flow or describe the relevant screen, document, or message step by step.
6. Close — 3 minutes
What part of this workflow have I misunderstood or missed?
Confirm any unfamiliar terms. Thank the participant and explain what will happen to the recording, notes, or artefacts.
If the workflow is narrow, a shorter session may be enough. If it spans several roles or systems, split it into multiple interviews instead of forcing a complete process into one call. See the guide to 15-, 30-, and 60-minute user interviews for choosing a realistic session length.
Turn the tour into evidence
After each conversation, keep the account traceable:
| Field | Capture |
|---|---|
| Workflow boundary | Where this tour started and ended |
| Typical stages | The participant’s high-level map |
| Specific episode | Date or context of the event reconstructed |
| Tools and artefacts | Systems, files, messages, notes, and physical objects |
| Handoffs | Who passed what to whom, and how readiness was signalled |
| Workarounds | Extra steps and what triggered them |
| Participant language | Verbatim terms, kept separate from your interpretation |
| Observation status | Reported, demonstrated, or directly observed |
| Evidence limits | Missing steps, sensitive areas skipped, and uncertain recall |
Compare tours within relevant roles and segments. A finance administrator’s workflow may differ from an approver’s even when both use the same product. Repeated patterns can guide a hypothesis, but an interview sample does not tell you how prevalent a problem is across the whole customer base.
Use product data, support records, usability testing, or a targeted survey when the next decision needs frequency, performance, or causal evidence.
How Maren supports workflow discovery
Maren’s UX Discovery style uses a grand tour to map the workflow, then follows with mini tours about tools, artefacts, handoffs, workarounds, and breakdowns. The conversation adapts to what each participant says rather than following a rigid script. Maren keeps the transcript, produces an interview summary, and can synthesise patterns across conversations.
Maren cannot watch a participant’s screen. Its evidence is still a participant’s account. A practical split is to use Maren to discover workflow patterns across more people, then run consented screen-share sessions for the few workflows where direct observation would change the decision.
That boundary matters. A clear spoken reconstruction can reveal a great deal, but it is not a substitute for seeing the extra tool, hidden spreadsheet, or repeated click in context.
Start with one workflow your product touches. Write one typical grand tour question and one specific version. Ask both before you mention the feature you are considering. The first useful discovery may not be a feature request at all. It may be a handoff, a workaround, or a word your product has never used.