Best for
- Founders about to interview users for the first time
- Replacing 'would you use this?' (everyone says yes) with questions that predict
- Getting consistent, comparable interviews across a research round
What you give it
- What you need to learn and who you will talk to
What you get back
- The script: opening, core questions in deliberate order, planned probes, close — timed to the slot
- Behaviour-first phrasing throughout: 'tell me about the last time…' replacing 'would you…' — with the reasoning taught inline
- The pitfall guard: your draft questions that lead, presume or pitch — flagged and rewritten
How it works
- Anchors every core question in past behaviour: the last time, the specific incident, the actual workflow — because memory of events predicts; speculation about futures flatters.
- Orders deliberately: context first, pain reconstruction next, your idea last — so their reality is on record before your pitch can reshape it.
- Plans the probes: the follow-ups written in advance (what happened next, who decided, what did it cost) so depth does not depend on live improvisation.
- Strips the poison: leading phrasings, embedded presumptions, and the pitch wearing a question costume — each flagged with why it corrupts.
Example
You: We are interviewing 12 clinic managers about appointment scheduling pain. Write the guide.
Result: The guide: a warm-up that establishes their actual current workflow (not their opinion of it), the core built on reconstruction — 'walk me through the last double-booking: what happened, who noticed, what did it cost' — with planned probes per question (silence, 'what happened next', 'who else was involved'), the pitch withheld until the final section (and framed as reaction-gathering, not validation), and four of your draft questions rewritten ('don't you find X frustrating?' led; 'how much would you pay?' predicts nothing — replaced with the current-spend reconstruction). Twelve comparable transcripts later, the synthesis wrote itself.
Limits — please read
- A script enables; interviewing is still a skill — the guide includes the conduct notes (silence tolerance, no selling) that matter most.
- Twelve interviews find patterns, not statistics; the guide's consistency is what makes the patterns comparable.
- What people DO still outranks what they SAY they did; where behaviour is observable, the guide says to observe it.