Customers Hire Products to Get a Job Done, Not to Buy Features客户雇用产品完成任务，而非购买功能

Jobs-to-be-Done reframes the unit of market analysis from who the customer is to what circumstance they're in — the same 40-year-old banker needs completely different things at a client dinner versus a Saturday breakfast with his kids.

The products that beat 'non-consumption' — Excel sheets, sticky notes, doing nothing — open the biggest blue oceans, because customers pay to get a job done, not for performance beyond what the job requires.

The argument's foundation is a claim about causality: behavior is caused by circumstance, not identity. Christensen's classic move — used here via the banker example — is to hold the person constant and vary the situation, showing the 'job' changes completely while every demographic label stays the same. This matters practically because most segmentation systems (CRM tags, ad targeting, loyalty tiers) still bucket customers by who they are, then recommend products based on that bucket. The piece argues this produces systematic mismatches: a 'high-net-worth male' tag tells a recommendation engine nothing about whether tonight's job is impressing a client or feeding a toddler.

The second move reframes what counts as a competitor. Conventional strategy compares your product to rivals in the same category; JTBD instead asks what job customers are currently solving without any dedicated product — via spreadsheets, sticky notes, or simply not solving it at all. The Salesforce-vs-Excel example is doing real work here: it explains why disruption theory and JTBD are natural partners — both locate growth in serving people the incumbents ignore, rather than out-featuring incumbents for people already well served. The corollary — customers won't pay for performance beyond the job's needs — is an implicit warning against feature bloat sold as differentiation.

The passive/active data split is the epistemological core of the piece. Passive data — clicks, sales, conversion rates — is cheap to collect precisely because it's a residue of past behavior stripped of context; that's also its weakness. Active data is the opposite trade: expensive to gather (requires observation and conversation), but it's the only source that carries the customer's actual struggle. The claim that ERP and CRM systems are 'systematically biased' toward passive data is the piece's sharpest structural insight — it's not that executives choose bad data, it's that the tools they're handed can only render one kind of truth, quietly reshaping what counts as evidence inside an organization.

Deep Growth ties the circumstance-based framing back to a business outcome: durable retention. The logic chain is — because jobs are tied to specific life circumstances, and circumstances recur, a product that nails an unmet job becomes load-bearing infrastructure in the customer's life rather than one option among many. This is a stickiness argument, not just a satisfaction argument: the moat isn't switching cost in the contractual sense, it's that the product has become part of how the customer routes around a recurring problem, which is harder to displace than a feature comparison chart would suggest.

The Big Hire/Little Hire split is the piece's most operational contribution because it locates where most retention strategies fail: teams over-invest in optimizing the purchase funnel and under-invest in the repeated, frictional moments after. The Airbnb case is chosen precisely because the two hires are physically separated in time and space — booking online versus standing at an unfamiliar door — making the risk of treating them as one continuous experience obvious. The prescriptive takeaway, mapping the customer's job timeline rather than just the purchase funnel, generalizes to any product with a gap between checkout and first real use — onboarding flows, assembly instructions, first-week engagement.

The Job Spec closes the piece by converting theory into an artifact a cross-functional team can actually build against. Its four components — functional thresholds, emotional floors, social signaling, and explicit trade-offs/obstacles — map onto the earlier active-data argument: these are exactly the dimensions passive data can't capture, so encoding them into a spec forces the 'thick data' collected earlier into a form R&D can act on. The morning-commuter milkshake trade-off (sacrificing flavor variety for convenience and duration) is doing double duty: it's a memorable illustration and a template for how any team should phrase its own trade-off clause.