A fast-food chain once spent months trying to sell more milkshakes. Thicker, sweeter, fruitier, sales didn't move. The data on who bought milkshakes was immaculate; the question itself was wrong. The fix wasn't a better milkshake. It was a better question: what job were people hiring the milkshake to do? That reframe is the whole of Jobs-to-be-Done, and it is one of the most useful research lenses a leader can carry into any decision about customers, products, or even their own team.

The quick version

  • People don't buy products; they hire them to make progress in a particular situation. That progress is the "job."
  • A job has functional, emotional and social dimensions, and it stays stable even as the products that serve it come and go.
  • You find jobs by studying the struggle that triggered a switch, not by asking people which features they'd like.
  • The move: interview people about a recent decision, map the forces pushing and pulling them, and design for the job, not the demographic.

The idea in depth: a job is progress, not a product

The phrase most people quote, "people don't want a quarter-inch drill, they want a quarter-inch hole", is usually credited to Harvard marketing professor Theodore Levitt, and sometimes to Clayton Christensen, who repeated it. Its real author was Leo McGivena, a New York newspaper executive, writing in the 1940s (the line surfaces in Percy Whiting's 1947 sales manual; the Quote Investigator traces the chain). The lineage matters less than the idea: a customer's goal is the outcome, and the product is just one disposable means to it. If a better means appears, they switch, and your beautifully engineered drill is suddenly furniture.

Clayton Christensen and his co-authors gave this a sharper frame in their 2016 Harvard Business Review article, "Know Your Customers' Jobs to Be Done", and the book that followed, Competing Against Luck (HarperBusiness, 2016). Their definition: a "job" is the progress a person is trying to make in a given circumstance. People "hire" products and services to get that job done, and "fire" them when something does it better. The milkshake study is theirs, the morning-commute buyers weren't hungry; they'd hired a thick, slow drink to make a boring drive bearable with one hand on the wheel. The competition wasn't other milkshakes. It was bananas, bagels and boredom.

The practical move: when you catch yourself listing features or sorting customers by age and income, stop and ask what progress sits underneath the purchase. The Christensen Institute frames a job along three dimensions, functional ("get a hole in this wall"), emotional ("feel competent, not foolish") and social ("look capable to my partner"). A demographic predicts none of these. Two forty-year-old fathers with identical incomes can be on completely different jobs in the same aisle.

flowchart LR
  A("Demographic lens
Who is the buyer?") --> B("Predicts almost nothing
about the choice")
  C("Jobs lens
What progress, in what situation?") --> D("Functional need")
  C --> E("Emotional need")
  C --> F("Social need")
  D --> G(["Design for the job"])
  E --> G
  F --> G
The shift from "who" to "what progress, in what situation." Leaders Loop

Two schools, and the limitation worth naming

Here is where honesty earns its keep, because Jobs-to-be-Done is not one settled method. There are two distinct schools, and they disagree about what a "job" even is.

The first is Christensen's jobs-as-progress view above: a job is the broad, emotionally-loaded goal you're striving toward. The second is Anthony Ulwick's Outcome-Driven Innovation, set out in his 2002 HBR article "Turn Customer Input into Innovation" (named one of HBR's breakthrough ideas of the year) and his book What Customers Want (2005). Ulwick treats the job as a precise functional process you can break into steps, then attaches measurable desired outcomes to each step, "minimise the time it takes to…", "reduce the likelihood that…", and prioritises which outcomes are important but poorly served. One school is a storytelling lens; the other is closer to a measurement system.

The limitation to name plainly: Christensen never gave a rigorous, testable definition of a job, which is exactly why the field splintered. The product strategist Alan Klement has documented these "two very different interpretations" and the confusion they cause when people use the same words to mean different things. JTBD is a powerful lens, but it is not a validated predictive instrument, treat it as a way to generate sharp hypotheses about needs, then test those hypotheses with harder methods. (For where those harder methods live, see qualitative vs quantitative vs mixed methods and validity, reliability & bias in research.)

So pick your school on purpose. If you're trying to find an unmet need or reframe a market, use the progress lens. If you're trying to improve an existing product against a known job, use Ulwick's outcomes, they're easier to score, prioritise and survey.

How you actually find a job: study the switch

You do not find jobs by asking "what would you like us to build?" People answer in features, and features are answers to a question they've already mis-framed. You find jobs by reconstructing a real, recent decision in forensic detail, what Bob Moesta, Christensen's collaborator, calls the switch interview. You rewind to the moment someone bought or quit something and walk the whole timeline: the first thought, the trigger, the alternatives weighed, the anxiety on the day.

Moesta's organising idea is the four forces of progress (discussed in his Intercom interview). Two forces push toward change, the push of the current frustration and the pull of the new option, and two forces hold it back: anxiety about the new thing and the sheer habit of the status quo. Habit, Moesta argues, is the strongest force of all; people keep doing what they do until something gets in the way or something better arrives. A purchase happens only when push and pull together overcome anxiety and habit.

flowchart LR
  P1("Push:
frustration with today") -->|toward change| S(["The switch"])
  P2("Pull:
attraction of the new") -->|toward change| S
  A1("Anxiety:
fear of the new") -.->|holds back| S
  A2("Habit:
comfort of the old") -.->|holds back| S
The four forces of progress: a switch only happens when push + pull beat anxiety + habit. Leaders Loop

In your next round of customer conversations, then, ban the words "want," "like" and "feature." Instead ask, "Take me back to the day you decided, what was going on?" Then listen for the four forces. The richest signal is the moment of struggle, not the list of preferences. The same lens works inside your own walls, too: when an employee resigns, the push and pull are usually obvious; the anxiety and habit you failed to address are the interesting part.

A worked example

Imagine a regional accountancy firm, call it Harbridge & Co, whose online booking tool is being quietly ignored. Clients keep phoning the front desk to book their year-end review instead. Leadership's instinct is a feature fix: a slicker calendar, SMS reminders, a nicer interface. (Figures here are illustrative.)

Before spending, the practice manager runs eight switch interviews with clients who booked by phone in the last month. The story that emerges has nothing to do with the calendar. The functional job is simply "get my tax sorted before the deadline." But the emotional job is "stop feeling anxious that I've missed something," and the social job is "look organised, not like the client who left it late." Phoning the desk does all three at once: a human voice confirms nothing is wrong, and they can ask the one nagging question that the booking form has no field for.

Mapped to the four forces, the push (deadline dread) and pull (a quick chat) both point to the phone. The online tool loses on anxiety, "what if I pick the wrong appointment type?", and habit. So the redesign changes: not a prettier calendar, but a booking flow that opens by reassuring ("You're on time, most reviews happen now") and ends with a free-text box, "Anything worrying you?", routed to a named adviser. Illustratively, online bookings climb from roughly 20% to 55% of the total in a quarter, not because the calendar got faster, but because it finally did the emotional and social jobs the phone call had been quietly carrying. The skimmer's takeaway: they weren't avoiding your tool; they were hiring the phone call to do a job your tool ignored.

Frequently asked questions

Is Jobs-to-be-Done just a fancy word for "customer needs"?

It's a specific way of framing needs. A "need" can be vague ("customers need convenience"); a job is anchored to a situation, an outcome and a moment of struggle ("when I'm mid-commute, help me stay occupied without making a mess"). That specificity is what makes it researchable, and stops "needs" from becoming a word everyone nods at and no one acts on.

How many interviews do I actually need?

Fewer than you'd think. JTBD practitioners often report that a small number of well-chosen switch interviews surface a handful of recurring patterns that cover most of a market. These are deep qualitative conversations, not a survey, but if you want to size a pattern across a population, that's a job for survey & sampling design, not more interviews.

Does this only apply to product and marketing?

No. The lens fits any moment someone "hires" something to make progress: an employee accepting a role, a manager adopting a new process, a board approving a strategy. Ask what job they're hiring it to do and which forces hold them back, and you'll predict adoption far better than a feature list or a mandate will.

Isn't asking people about the past unreliable, they'll just rationalise?

Memory is imperfect, which is why the switch interview hunts for concrete, time-stamped detail ("what did you do right before?") rather than opinions, and why you triangulate across several people for the same pattern. It's a hypothesis-generator, not proof. Treat its output as claims to test, and mind the usual interview biases.

Christensen or Ulwick, which should I use?

Use Christensen's progress lens to discover or reframe a need and find white space; use Ulwick's Outcome-Driven Innovation to measure and prioritise needs against a product you already have. Many teams discover with one and prioritise with the other.

Related in the Toolkit

Where to go next