Jobs-to-be-done is the most useful strategy lens in product management and the most frequently reduced to a slogan. Teams put "people don't buy drills, they buy holes" on a slide, rename their personas "jobs," and change nothing. Actually practicing JTBD means doing forensic interviews about real past decisions — and letting the answers redraw your competitive map, which is the part teams flinch from.
What JTBD looks like when it's real
At WisOwl AI, the assumed job was "help recruiters find candidates," with other hiring platforms as the competition. JTBD interviews — walking recruiters through the last three roles they actually filled, hour by hour — showed the real hiring workflow ran on WhatsApp groups, personal networks, and a spreadsheet. The job wasn't search; it was "fill this role this week without staking my reputation on an unknown." Our semantic matching engine mattered less than trust signals and speed-to-shortlist. We rebuilt onboarding and agent behavior around that, and it's a large part of how the platform reached 5,000+ signups and 15+ recruiter partnerships organically.
At CaaStle, the same lens applied to subscription commerce: customers weren't hiring a clothing subscription to own more clothes — many were hiring it for the recurring feeling of newness with a bounded budget. That framing shaped retention features in ways a demographic persona never could have.
What an engagement includes
- Switch interviews: 10–15 forensic conversations about real purchase and churn decisions — what triggered the search, what almost stopped it, what the product replaced.
- Job map and forces analysis: the progress your customers are trying to make, and the anxieties and habits holding them in place.
- Strategy translation: the part most JTBD work skips — turning the job into positioning, onboarding design, roadmap bets, and a competitive map that includes the spreadsheets and WhatsApp groups you're actually up against.
Four to six weeks, and it pairs unusually well with a GTM or PMF engagement, since all three are ultimately the same question: what is your product hired to do, and how strong is the evidence?