Jobs to Be Done

In brief

Jobs to Be Done (JTBD) is a theory of why customers buy. Its central claim is deceptively plain: no one wants a product. People want to make progress in their lives, and they hire products, and sometimes services, habits, or workarounds, to get that progress done. Clayton Christensen, the Harvard Business School professor best known for disruptive innovation, put it most sharply in Competing Against Luck (2016): the job, not the product and not the customer, is the correct unit of analysis for a market. Get the job right and you can predict what people will buy, because you understand the causal reason behind the purchase. Get it wrong, and you are left correlating traits (age, income, product category) that describe buyers without explaining them.

The full treatment

The problem it answers

Most market research explains almost nothing. It tells you that a certain product sells better among 25-to-34-year-old suburban homeowners, or that customers who bought X also bought Y. These are correlations. They describe who buys and what they buy alongside what, but they do not say why, and correlation is a poor guide to what to build next. Christensen's complaint was that companies drown in data about products and customers and still cannot reliably create products people want. Innovation, in his phrase, was left to luck. JTBD is his attempt to supply the missing ingredient: a causal theory of demand, an account of the mechanism that actually produces a purchase.

How it works

The unit of analysis is the "job to be done." A job is the progress a person is trying to make in a particular circumstance. Note the two parts. First, it is progress, not a task in isolation: it has functional, emotional, and social dimensions all at once. Second, it is anchored in a circumstance: the same person has different jobs at different moments, and the circumstance, not the person's stable traits, is what fixes the job. When a job arises, the customer looks around for something to hire to do it. If the hire performs, they hire it again; if it disappoints, they fire it and look for a replacement, which may not be another product in the same category at all.

The idea inverts the usual question. Instead of asking "what features should this product have," you ask "what job is the customer hiring this product to do, and what are they hiring instead of it." That reframes the competitive field. A tax-software company is not only competing with other tax software; it may be competing with a human accountant, a spreadsheet, or the customer's decision to do nothing. Competition is defined by the job, not by the product category.

The milkshake example

The story Christensen used to make this stick, drawn from consulting work he recounts in Competing Against Luck, concerns a fast-food chain trying to sell more milkshakes. The company had already done the obvious thing: it profiled the typical milkshake buyer, convened panels of such customers, and asked them how to improve the shake (thicker, chunkier, cheaper, more chocolatey). The tweaks changed nothing.

The JTBD move was to ignore the customer profile and watch the circumstance. Researchers logged when milkshakes were bought and found a large share sold before 9 a.m. to solo commuters who took the shake with them. Interviewed, these buyers revealed the job: the drive to work was long and dull, they were not yet hungry but would be by mid-morning, and their free hand needed something to do. They had "hired" a range of things for this job and fired them all. A banana was gone too fast. A doughnut left greasy fingers and crumbs. A bagel was dry and demanded two hands. The milkshake won because it was viscous enough to last the whole commute through a thin straw, tidy, and filling. It was competing with bananas and boredom, not with other milkshakes. That insight pointed to real improvements (make it thicker, add chunks of fruit to reward the straw, move it to a pre-pay counter for a faster grab) that no amount of asking "how do we make a better milkshake" would have surfaced. A separate afternoon job, a parent placating a child, turned out to want almost the opposite product.

Distinctions that matter

JTBD is not the same as gathering customer requirements. Customers are experts on their struggle, not on the solution; asking them to specify features tends to yield faster horses. JTBD is also not market segmentation by demographics. The unit is the job, and a single job cuts across ages, incomes, and categories. Finally, the "job" is not merely functional. Christensen insisted the emotional and social dimensions (how the hire makes me feel, how it makes me look) are part of the job, which is why a purely spec-sheet reading of the theory misses it.

Lineage

The oldest root is a line usually credited to the Harvard marketing professor Theodore Levitt: people do not want a quarter-inch drill, they want a quarter-inch hole. Levitt's 1960 essay "Marketing Myopia" made the same structural point, that firms fail when they define themselves by their product rather than by the customer need they serve. Christensen's contribution was to turn that aphorism into an operating theory with a defined unit of analysis. He first stated the job-to-be-done framing formally in The Innovator's Solution (2003), written with Michael Raynor, as the natural next question after disruptive innovation: once you accept that markets are unstable and get remade from below, you need a stable way to see demand, and the job supplies it. He and his coauthors gave it book-length form in Competing Against Luck (2016). A parallel and more measurement-driven strand, Anthony Ulwick's Outcome-Driven Innovation (from 2002), arrived at a compatible view by breaking jobs into quantifiable desired outcomes. The theory shares intellectual air with product-market fit: the job is one precise answer to what "fit" is fit with.

The strongest case for it

The framework earns its keep by reframing competition and directing attention where causation lives. By insisting on the circumstance, it explains anomalies that demographic segmentation cannot, and it repeatedly points builders toward improvements they would otherwise never see, as the milkshake case shows. It guards against the classic innovation trap of adding features that focus groups request but no one hires the product for. It also widens the field of vision in a commercially useful way: seeing that your real rival is the customer doing nothing, or using a spreadsheet, or hiring an accountant, is often more valuable than benchmarking against the named competitor. And it is portable. The same lens works for a consumer good, a business tool, a hospital service, or a nonprofit, because every purchase, in this view, is a hire against a job.

The strongest case against it

The most serious charge is that JTBD is hard to falsify. A "job" is identified after the fact from how customers behave, and almost any purchase can be narrated as the hiring of a product for some job, so the theory risks explaining everything and predicting nothing in advance. Critics note that Christensen never specified a rigorous procedure for isolating the "true" job or for telling a good job-definition from a bad one, which leaves the method dependent on the interviewer's skill and open to just-so storytelling. This is a cousin of the empirical scrutiny disruptive innovation has drawn: Andrew A. King and Baljir Baatartogtokh argued in 2015 that even Christensen's flagship theory fit far fewer of its own case studies than claimed, and skeptics extend the worry that his frameworks are compelling narratives more than tested theories.

There is also a schism inside the movement. One camp, associated with Ulwick, treats a job as a stable set of measurable outcomes; another, associated with the "switch" interviews of Bob Moesta and Chris Spiek, treats a job as the emotional story of a moment of purchase. The two schools disagree about what a job even is, which undercuts the claim that JTBD is a single settled unit of analysis. Finally, the framework says little about supply, price, distribution, or competitive dynamics; it clarifies why a customer might hire you but not whether you can build, fund, or defend the offering, so it is a lens on demand rather than a full theory of a business.

Where it stands now

JTBD is entrenched in product management, design, and startup practice, taught in business schools and built into how many product teams frame their work. The "switch" interview and the job story ("when [situation], I want to [motivation], so I can [outcome]") are standard tools. Its influence outran its academic validation: it lives more as a durable practitioner heuristic than as a formally tested economic theory, and the internal split over what a job is has never fully closed. That combination, wide adoption and thin proof, is common to Christensen's ideas. What is not in dispute is that the reframing stuck. Asking "what job is the customer hiring this for" has become one of the default first questions in the building of new products.

Test yourself

Pick something you bought recently that you did not strictly need. Set aside the features and your own demographic profile and ask: what was the circumstance, and what progress were you trying to make? Now name what else you considered hiring for that same job, including doing nothing. If the honest answer sits in a different product category than the thing you bought, you have found a job, and probably a market someone is not yet serving.