Competing Against Luck

The central claim is that correlation about customers — age, income, role, segment — is not causation about purchases. Knowing a great deal about a buyer still leaves you guessing about what they will do.

The unit of analysis Christensen proposes instead is the *job*: the progress a person is trying to make in a specific circumstance. The job, not the customer, is what causes a purchase.

Customers do not buy products. They hire them to make progress in a circumstance.

Devmark’s reading — not a quotation from the book

A job is much more specific than a need, and the specificity is the point. It has a circumstance attached, it is functional, emotional and social at once, and it persists over time even as the products hired for it change.

"I need convenient food" is a need and predicts nothing. "I have a long dull commute alone and want something to occupy me that will keep hunger off until eleven" is a job, and it rules most products out.

The book's most retold example, from research into why people bought milkshakes at a fast-food chain. A large share were bought early in the morning by solo drivers.

The finding was that the shake was competing not with other shakes but with bagels, bananas and doughnuts for the job of making a dull commute tolerable — and that it won because it was thick enough to last, one-handed and non-messy. Product improvements aimed at "shake buyers" would have missed all of that.

Once the job is the unit, the competitive set changes entirely. Your real competitor is often not a rival product but the customer doing nothing, or doing something crude that has nothing to do with your category.

This is the framework's most commercially useful consequence: markets look much larger and much stranger when measured by the job rather than by the category.

Christensen argues progress is blocked in identifiable ways, and that the richest research is into the moment of switching — what happened just before someone changed what they hired.

His forces model is genuinely useful: the push of the current situation and the pull of the new solution work in favour of change, while anxiety about the new and attachment to the current habit work against it. Most product teams work only on the pull and ignore the two forces of resistance.

Everyone works on making the new thing more attractive. Almost nobody works on the anxiety about switching.

Devmark’s reading — not a quotation from the book

The organisational half of the argument: a company that understands a job must organise its processes around delivering it, not around functions or product lines.

Christensen's claim is that this is where durable advantage sits, because competitors can copy a product far more easily than they can reorganise to deliver an experience that resolves a job completely.

His sharpest organisational warning: as a company scales, the job gets displaced by data about the job, and then by data that is merely available.

He calls this the fallacy of active versus passive data and the fallacy of surface growth. Both describe an organisation gradually optimising for numbers that were once proxies for the job and are now untethered from it.

The main criticism is that jobs are identified retrospectively and there is no reliable method for finding them prospectively. The milkshake story is compelling and is also a story about a job that was found by someone looking hard, which is not a procedure.

Critics also note that the framework is close to unfalsifiable: any purchase can be described post hoc as hiring something for a job, which makes the theory a lens rather than a predictive instrument.

This is Christensen's attempt to answer the question his earlier work raised but did not solve. *The Innovator's Dilemma* explains why incumbents lose; it does not tell anyone what to build.

Jobs to Be Done is the constructive half — a theory of where to look for the innovation that will actually be adopted. Read in that sequence the two books fit together properly.

The interview discipline is the most portable part: research the switch, not the preference. Ask what happened in the days before someone changed, what they tried first, what nearly stopped them.

And the forces model, which converts vague "customers aren't adopting" into four separately addressable things — push, pull, anxiety and habit — of which most teams are only working on one.