gardrutdal treats each interview as a single case. A stack of cases can suggest a pattern. It cannot stand in for a census of a market, and it cannot support a forecast of savings. The words to keep are the buyer’s: who spoke, what almost stopped the purchase, which proof was demanded, and which phrase is still used in internal meetings months later.
The buying group is a set of jobs
A persona poster collapses those jobs into one friendly character. Claude listening refuses that collapse. The economic buyer may never open the product. The champion may overstate consensus because their reputation is now attached to the choice. The user may have wanted a different tool and lost. The blocker is often in security, legal, finance, or an adjacent plant that will be forced onto the same login. Procurement may be copying last year’s scoring sheet. Definitions of these seats live in the glossary.
Interviewing only the champion produces a flattering story. The champion was in the room when the decision felt inevitable. They are a poor witness to the week it nearly died. Pair the champion with someone who felt the friction in their own calendar — a pattern described further under purchase friction. Disagreement between the two memos is a finding, not a defect to be smoothed before the read-out.
What the hour is for
Seventy-five minutes is enough to rebuild one purchase and not enough to tour a roadmap. Ask for the last time this category was bought or renewed. Stay on that instance. You want the week it became someone’s problem, the alternative that was nearly chosen, the meeting where the file stalled, and the artifact that unblocked it: a security questionnaire, a reference call, a pilot on one line, a redline from counsel. You do not ask whether they would pay more. That question invites a polite fiction and contaminates later positioning synthesis.
People remember status and sequence better than they remember abstract importance. “Security was critical” is a slogan. “The questionnaire sat with a contractor for eleven days and the plant manager started emailing the incumbent again” is evidence. Collect the second kind.
A question line that stays on their side of the table
- Walk through the last time this category was bought or renewed, starting from the week it became someone’s problem.
- Who had to say yes, and who could have stopped it without formally voting?
- What did you almost choose, and what made that alternative lose?
- Where did the process sit still, and what was the room actually waiting on?
- Which words do you use internally — not the words vendors use on their sites?
- What would a skeptical colleague say is still unproven?
The craft of probes, silence, and recording consent is written up separately in interview design. Do not improvise a demo in the gap where a probe should go. If they ask what you sell, answer in one sentence and return to their timeline.
The same-day memo
Write before the next call repaints the hour. Four headings are enough. Scene: where you were and what role they actually hold, as distinct from the title on the invitation. Verbs: the actions they described, in their order. Stall: the concrete wait. Negative note: anything that failed to fit the story you hoped to hear, including a purchase that never should have been hard and was easy for a dull reason. Those negative notes are how pattern finding stays smaller than a slogan.
Quote short phrases. Paraphrase the rest and mark it as paraphrase. Do not upgrade “we muddled through the security review” into “buyers demand enterprise-grade trust.” The first can be checked against a note. The second cannot.
What you must not conclude
One interview does not reveal a segment. Five interviews do not reveal a market share. A story about a avoided outage does not reveal a payback period. Claude fieldwork ends the interview in a memo, not in a slide that says the market has spoken. If a later positioning line cannot point back to a stall, a verb, or a negative note, it does not belong on the page. Observation of the places those stalls occur is the companion practice in market observation.