A discussion guide is not a list of questions. It is an order of questions, built so that each answer makes the next one askable. The most common failure is opening with the expert's background, which spends the calibration window on information you already had.
What is a discussion guide?
An order of questions with a purpose behind the ordering. A list of questions is what you write when you know what you want to ask; a guide is what you write when you know what you want to find out, which is a different problem.
The distinction shows in what happens when an expert answers question four while answering question two. A list forces you to ask four anyway; a guide tells you what to ask instead — and carries the segment question that resolves two experts who appear to disagree.
- 01
Write the decision
State the decision the research serves, not the topic it covers.
- 02
Open on calibration
Lead with a specific, partly checkable question rather than with background.
- 03
Order observed to interpretive
Ask what they saw before asking what they think it means.
- 04
Reserve a third
Leave roughly twenty minutes unscripted for what you did not know to ask.
What goes in the first ten minutes?
Something specific and checkable that calibrates how current their knowledge is. Not their background — that is in the profile you approved before the call was booked — but a question whose answer you can partly verify, which tells you how much weight to put on everything after it.
Opening with credentials wastes the one window in which an expert is most willing to be precise, and it signals that the call is a formality.
How should questions be ordered?
From what the expert observed toward what they interpret. Observations are recoverable and checkable; interpretations are where the value is and they are only worth having once you know what they rest on.
Asking for the interpretation first tends to produce a position the expert then defends for the remaining fifty minutes, including against their own observations.
| Question list | Discussion guide | |
|---|---|---|
| Built from | What you want to ask | What you want to find out |
| When an answer arrives early | You ask the question anyway | You move to what it makes askable |
| Across a program | Comparable and shallow | Comparable on the spine, deep on the probes |
What happens when they go off-script?
Usually you follow. The unplanned direction is where findings come from that nobody could have specified in advance, which is the entire reason for talking to a person rather than reading a report.
The guide should reserve time for this rather than fill it. A sixty-minute call supports about eight real questions with follow-ups; guides with twenty become surveys read aloud.
Should you reuse a guide?
Keep the spine and vary the probes. A fixed spine makes answers comparable across a program, which is what lets you say anything about the set rather than about individuals — and what makes forty interviews in three weeks readable at all.
Fixed probes do the opposite. Different experts need different routes to the same point, and asking everyone the identical follow-up produces the same shallow answer from all of them.
Reuse also has a limit worth respecting. A spine written for one market carries assumptions about how that market works, and applying it unchanged to another produces questions that make sense to the researcher and not to the expert. The spine travels; the framing around it usually does not.