Last updated: July 28, 2026
\n\nJobs-to-Be-Done interviews uncover the real reason someone \"hired\" a product by reconstructing the timeline of their decision — the moment they switched. Instead of asking what features people want, you interview 5–10 recent adopters about the circumstances, struggles, and forces that pushed them from an old solution to a new one. The output is the underlying job, not a feature wish list.
\n\nJobs-to-Be-Done interviews are qualitative interviews designed to surface the \"job\" a customer is trying to get done when they buy or adopt a product. The framing comes from Harvard Business School professor Clayton Christensen's Jobs-to-Be-Done theory: people don't buy products, they \"hire\" them to make progress in a particular circumstance. The interview's job is to recover that circumstance in enough detail that the demand becomes explainable and, ideally, predictable.
\nTwo schools shape modern practice. Tony Ulwick's Outcome-Driven Innovation treats the job as a set of measurable desired outcomes. Bob Moesta and Chris Spiek's Switch interview treats the job as a story with a timeline — the approach most user researchers and product teams reach for, and the focus of this guide.
\n\nStandard discovery questions — \"What features would you like?\" \"How likely are you to use X?\" — ask customers to predict their own future behavior and design your product for you. Both are things people are bad at. JTBD interviews sidestep this by anchoring on something people are good at recalling: what actually happened the last time they changed solutions.
\nThat shift matters because it changes what you can act on. A feature list tells you what to build next quarter; a well-understood job tells you what problem your roadmap is actually for. It also travels well across our other guidance on writing non-leading interview questions and continuous discovery.
\n\nA Switch interview walks backward and forward through one real purchase. You are not collecting general attitudes; you are documenting a single timeline in forensic detail — when the first thought struck, what triggered the search, what nearly stopped them, and what finally tipped the decision.
\n\nUse these six phases as the backbone of your guide. Ask for concrete events and timestamps (\"What day was that?\" \"Where were you?\"), not summaries.
\n| Phase | What you're recovering | Example prompt |
|---|---|---|
| 1. First thought | When the old solution first felt inadequate | \"Take me back to the first moment you thought there had to be a better way.\" |
| 2. Passive looking | Casual, low-energy awareness of alternatives | \"Between then and actually searching, what did you notice or half-consider?\" |
| 3. Active looking | The trigger that made the search intentional | \"What happened that made you actually start looking?\" |
| 4. Deciding | The trade-offs and the final tipping point | \"What almost stopped you? What tipped it in the end?\" |
| 5. First use | Whether the first experience matched the hope | \"The first time you used it, what surprised you?\" |
| 6. Ongoing progress | Whether the job is actually getting done now | \"What can you do now that you couldn't before?\" |
The two highest-energy moments — the first thought and the moment of committing — are where the real forces reveal themselves. Slow down there.
\n\nEvery switch is a contest between forces that promote change and forces that resist it. Moesta's Four Forces model gives you the analysis lens:
\nA switch only happens when push + pull outweigh anxiety + habit. This is strategically useful: most teams over-invest in pull (more features, louder marketing) when the real blocker is anxiety (onboarding fear) or habit (switching cost). JTBD interviews tell you which force is actually binding.
\n\nInterview people who adopted or switched in the last three to six months — recent enough that the decision is vivid, distant enough that the job is real. Five to ten interviews is a common starting point for a single job; watch for saturation rather than fixing a number in advance.
\nWhen a participant offers an opinion (\"I like clean interfaces\"), redirect to the event (\"Tell me about the moment you decided\"). Follow energy: hesitation, contradiction, and emotion mark the forces you're hunting.
\nTag each transcript segment by force (push / pull / anxiety / habit) and by the progress the person was trying to make. Cluster across interviews to name the recurring job. This is qualitative coding applied to a specific frame.
\n\nBefore you field a Switch-interview study, confirm you can answer yes to each:
\nQualitati is an AI user research platform for product, UX, and customer-insights teams — and JTBD interviews map cleanly onto what it automates.
\nThe AI accelerates the mechanical work — running the timeline, tagging forces, keeping every code tied to its quote. Naming the job, and deciding which force your roadmap should attack, stays with the researcher.
\n\nJTBD interviews are powerful but not universal, and it's worth being honest about where they strain.
\nHuman-review note: JTBD framing choices (what counts as \"the job,\" how forces are coded) shape conclusions — validate them against your team and, for consequential decisions, a research methodologist rather than tool output alone.
\n\nFive to ten recent switchers is a common starting point for a single job. Keep going until additional interviews stop changing your understanding of the job — the standard saturation logic in qualitative research.
\nA normal user interview often gathers attitudes, preferences, and feature requests. A JTBD (Switch) interview reconstructs one real purchase as a timeline to find the causal forces behind the decision — it studies behavior and circumstance rather than opinion.
\nPush of the situation and pull of the new solution promote change; anxiety about the new and habit of the present resist it. A switch happens only when push and pull outweigh anxiety and habit.
\nPeople who adopted your product or switched from an alternative in roughly the last three to six months, so the decision is still fresh and specific rather than generalized.
\nYes. An AI moderator can follow the Switch-interview timeline and probe missed moments at scale, and AI-assisted coding can tag the Four Forces across transcripts. Human researchers still name the job and interpret the emotionally charged decisions.
\n\nJobs-to-Be-Done interviews replace guesswork about what to build with evidence about why customers switch. Run 5–10 Switch interviews with recent adopters, walk the timeline, code for the Four Forces, and let the recurring job — not a feature list — guide the roadmap. With an AI user research platform, you can run those interviews and code the forces at a scale manual work can't reach, while keeping interpretation human.
\nStart free with 30 credits — no credit card required. Create an account, view transparent pricing, or run an AI-moderated interview, focus group, conversational survey, or thematic-analysis project with Qualitati.
\n","related":[{"slug":"screener-ai-moderated-interviews-2026","title":"How to Write a Screener for AI-Moderated Interviews (2026)","description":"A 2026 template for AI-moderated interview screeners that filter for fit and resist participant fraud, with a signal table and scoring rubric.","category":"How-To","date":"2026-06-16","readTime":"10 min read","author":"Qualitati Research Team"}]}