User interviewing
User interviewing is live conversation with users, typically a phone or video call, aimed at learning what they were trying to accomplish in their own lives that brought them to the product. The interviewer's job is overwhelmingly to listen: the user's exact words, the order they remember things in, the connections they draw between things the interviewer wouldn't have linked. That's the data, which is why the discipline is to extract what's already in the user's head rather than pitch hypotheses for them to react to or steer the conversation onto product feedback.
Definition
User interviewing is the live-conversation method within user value research — one-on-one data-gathering with users about their transition, structured by the situation explorer and governed by listen, don't talk.
Sibling methods
Email inquiry and starting-churn CS ticket pattern analysis are the other two methods feeding intent clustering. Each produces a different data shape: written narrative from email replies, in-pursuit articulations from CS tickets, and live conversational data from interviews. Interviews typically follow email inquiry when replies surface patterns warranting deeper exploration — the reply itself is the foot-in-door for the conversation.
Interview structure
The situation explorer supplies the structure: a timeline-organized grid covering sought state, prior states, route, what comes after, and completion criteria. The interviewer fills the sought-state column first because its conditions define the requirement-frame everything else is in service of producing. The worksheet supplies regions the interviewer knows they'll explore, not pre-fabricated questions; questions emerge from the shape of what the interviewee is describing within whatever region the conversation is currently in.
Listen, don't talk
The discipline holds because the interviewee's mind is the source of data, and only listening accesses it. The interviewee's exact articulation is what the method captures — how they word their situation, the sequence in which they remember things, the connections they make. Interruptions, redirections, mid-stream verifications, and pre-scripted questions impose the interviewer's framing on that content and corrupt the signal.
Transition focus is the paired discipline. The interviewer asks about what users want in their lives, treating the offering as one source of affordances among many they're marshaling toward an outcome. Offering-feedback questions ("what would you change about feature X?") route the conversation onto the company's timeline rather than the user's, which is the wrong stack.
Who gets recruited
Recruitment depends on the investigation question. The canonical application is where a multiplier analysis finding identifies a population with a specific intent. Subjects come from that cohort — users who performed the candidate behavior, reached the commercial outcome event, and did so within the conversion window. Interviewing a broader population dilutes the signal across multiple intents.
Other questions call for other patterns. Churned customers for failure-mode questions; sustained customers for long-arc value questions; missed conversions for recovery-opportunity questions; category users broadly for offering-positioning questions. Each population stands at a different point in the user-provider trajectory and produces different signal.
Example
A SaaS analytics team finds via multiplier analysis that users who connect a data source in their first session convert at 4× the base rate. They recruit eight subjects from that cohort and run interviews using the situation explorer. The sought-state column fills with a consistent pattern: variants of "needed retention numbers in a board deck due Friday." Working backward through route and prior state, what mattered turns out to be speed-to-first-chart, not data-source connection in particular — the connection step was a means to that end and would have been substitutable with any path that hit the same speed. The intent theory becomes "deadline-driven IC pushing a number to leadership" rather than "team adopting analytics tooling."