Jobs To Be Done
A lens that frames innovation around the progress a person is trying to make — the job they hire a product to do — rather than the product’s features or the customer’s demographics.
People do not buy products. They hire them to do a job, and they fire them when something does it better. Understand the job and you understand the demand.
What it is
Stop describing the customer. Describe the job.
Jobs To Be Done is a way of understanding customer needs by focusing on the progress a person is trying to make in a particular circumstance — the “job” they are trying to get done — rather than on the product’s attributes or the customer’s profile. The core idea, associated with Clayton Christensen, is that people do not buy products; they hire them to do a job, and they fire them when something does it better. Understanding the job — including its functional, emotional, and social dimensions — reveals why people really make the choices they do, and where the genuine opportunities for innovation lie.
Its distinctive move is to look past two things most analysis fixates on: product features and customer demographics. Neither reliably predicts behavior. A demographic tells you who bought the product; the job tells you why they hired it. A feature tells you what the product does; the job tells you what progress the person was trying to make. Expressed as a job — “when I am commuting alone and bored, I want something engaging and filling I can manage one-handed, so I can arrive at work feeling ready” — the same need makes the true competitors and the real improvement immediately visible. Neither the demographic nor the feature would have revealed either.
The job is expressed as progress in a circumstance: when [situation], I want to [motivation], so I can [outcome]. Note what is absent: no product, no demographic, only the situation, the desired progress, and the outcome. That is the reframe, and it is the whole method.
Build the job statement
Stop describing the product. Build the job. Click the slots.
Toggle between the wrong framing (a feature/demographic) and the right one (a job statement) to feel the reframe. In the job view, click each slot to reveal what belongs there and how to get it right.
When to deploy it
A lens for understanding demand — not a source of data.
Use JTBD when
- →You want to reframe what you are really solving for — past features and past demographics — in terms of the progress people are trying to make.
- →You suspect the team is too focused on product attributes or demographic segments that do not actually predict behavior or drive switching decisions.
- →You need to find unmet needs, or to see the true competition, which is often not what the product category implies.
- →You are trying to understand why people switch to or from a solution — the moment of hiring and firing, which is where the job is most legible.
Do not lean on it when
- ×You need an execution or build process. JTBD is a lens for understanding demand, not a method for building. It reframes the question; it does not answer it.
- ×You have no real research into the circumstances of use. A job written from assumption is just a feature or a guess dressed as a job — and it carries the same false confidence.
- ×The job is already well understood and the decision genuinely is about a feature or a segment. Not every problem needs reframing.
The honest limit: JTBD is a lens, not a source of data. It reframes understanding you already have into terms of progress; it cannot substitute for the research into real circumstances of use that a true job statement depends on. Its most common failure mode is a “job” that is secretly a product feature or a demographic in disguise.
How it works
Five moves, from circumstance to demand.
Investigate the circumstances of use — especially the switch.
Study when and why people hire a solution, paying special attention to the moment of switching: what triggered them to seek something new, what they used before, and what progress they were chasing. The switch is where the job is most visible. It reveals the unmet demand that finally became acute enough to produce a behavior change.
Articulate the job as progress in a circumstance.
Write the job independent of any solution: "when [situation], I want to [motivation], so I can [outcome]." If a product name or a demographic appears in the statement, it is not yet a job — keep rewriting until only progress-in-circumstance remains. The test: could a completely different product satisfy this job statement? If yes, you have a job.
Surface all three dimensions.
Identify the functional (the practical task), emotional (how the person wants to feel), and social (how they want to be perceived) dimensions of the job. The emotional and social dimensions are usually where the non-obvious insight lives — and where the teams that see only the functional job miss most of the real demand.
Find the real competition.
With the job defined, ask what else people hire to do this job. The true competitors are often nothing like your product — the milkshake competes with bananas and boredom, not other milkshakes. This cross-category competition is one of JTBD's biggest strategic payoffs, and it is invisible until the job is defined.
Reframe the problem and the opportunity.
Use the job to redirect the team from improving the product's features to serving the job better. This often points to entirely different improvements than a feature-focused view would — the thicker milkshake that lasts the commute, not the cheaper one that competes on dessert value.
Best practices
What good looks like — and what prevents it.
When it goes well
- ✓The job is defined in terms of progress and circumstance — no product name, no demographic, only the situation and the desired outcome.
- ✓The emotional and social dimensions are surfaced alongside the functional one — the job includes how the person wants to feel and to be perceived, not just what they want to accomplish.
- ✓The framing reveals unexpected competition: things people actually hire to do this job that look nothing like your product.
- ✓It is grounded in real investigation of the circumstances of use — especially the switching moment, where the job is most clearly legible.
- ✓It redirects the team from feature-polishing to job-serving, pointing at improvements that a feature-focused view never would have reached.
The mistakes, and how to avoid them
Writing the job as a product feature in disguise.
"The job is a thicker milkshake" is not a job; it is a feature. If a solution appears in the statement, keep rewriting until only progress-in-circumstance remains. The test: remove your product from the statement. Is the need still clear? If not, a solution crept in.
Only capturing the functional dimension.
The practical task surfaces easily; the emotional and social dimensions take deliberate effort. Build in a specific step to ask how the person wants to feel and to be perceived — or the team will stop at the functional layer and miss most of the job.
Confusing the job with the demographic.
"The job is what suburban parents want" collapses back into segmentation. The circumstance predicts behavior; the demographic does not. Keep the person's profile out of the job statement and keep the situational trigger in.
Skipping the switching moment.
The moment someone hires a new solution or fires an old one is where the job is clearest and most specific. Analysis that ignores the switch misses the strongest evidence and often produces a job that is too generic to reveal real competition or real improvement.
Logistics
Running the investigation from switching moment to job statement.
JTBD is primarily a research and synthesis orientation, not a workshop exercise. The richest evidence comes from interviewing people about a specific recent switch — and the interview technique matters. These are some of the practical things that determine whether the investigation finds the real job.
Research the switch directly
The richest JTBD evidence comes from interviewing people about a recent switch: when and why they started using something new, what they had before, what finally pushed them. Structure the research around the switching moment rather than around satisfaction with the current product. 'Walk me through the last time you changed how you were doing this' is more revealing than 'how satisfied are you with our product?'
Interview for circumstance, not opinion
JTBD research works by getting people to narrate specific recent episodes, not by asking what they want. 'Walk me through the last time you...' reveals the job; 'what would you like us to improve?' reveals stated preferences, which are much weaker evidence. The job hides in the story of a real circumstance.
Push past the functional job in the room
The functional job surfaces easily and tends to dominate early synthesis. The emotional and social dimensions take deliberate probing. Build in a specific prompt to ask how the person wanted to feel during and after the job, and how they wanted to be perceived — or the team will stop at the functional layer and produce a job statement that looks complete but misses the job's deepest drivers.
Watch for solutions sneaking into the job
As the team writes job statements, actively police them for product names, features, and demographics. A quick test: could a completely different product also satisfy this job statement? If not, a solution has crept in. A job that passes the test opens the space to see real competition and real alternatives.
Map the real competition explicitly
Once the job is defined, run a short exercise listing everything people hire to do this job — however unlike your product. This is where the strategic surprise and the real market definition usually appear. The non-obvious competition is often the most revealing output of the whole investigation.
AI and this method
AI writes clean job statements instantly. The real job is usually the surprising one it does not reach.
Toggle between modes to see what AI produces from a product description — and where the counterintuitive real job, the emotional dimensions, and the non-obvious competition stay out of reach.
In-depth example
The same product. Two investigations. One found the job; one did not.
A fast-food chain wants to sell more milkshakes. Sales are flat. Both approaches investigate the same product and the same purchase; only the method differs. Toggle between them to see which one found the job — and what the job revealed.
Scenario
A fast-food chain wants to sell more milkshakes. Sales are flat.
Conventional analysis had already failed. The team had segmented by demographics, asked customers how to improve the shake — thicker, cheaper, chunkier — and made none of those changes. Nothing moved. They switched to a JTBD investigation: study the circumstances of purchase, especially the moment of hiring.
What the investigation revealed
A striking number of milkshakes were bought early in the morning, by solo commuters, alone, buying nothing else. Investigating that specific circumstance revealed the job: make a long, boring solo commute more interesting, and stave off hunger until lunch — all manageable with one hand on the wheel.
That job had almost nothing to do with milkshakes as food. It was a companionship-and-engagement job in disguise — and the milkshake happened to do it well: thick enough to last the whole commute, one handle, engagingly dense to consume.
The job statement
WHEN
I face a long, boring solo commute
I WANT TO
something engaging and filling I can manage one-handed
SO I CAN
arrive at work feeling ready, not depleted or bored
No product name. No demographic. Only progress-in-a-circumstance — which is exactly why the real competition and the real improvement suddenly become visible.
The real competition (not other milkshakes)
Banana
Gone too fast — leaves you hungry by 9am
Bagel
Dry, needs two hands, messy at the wheel
Donut
Messy, leaves fingers sticky for the meeting
Boredom
Nothing — you just sit there, drained
The winning move — revealed by the job
Make it thicker so it lasts the whole commute (the opposite of what the feature-improvement analysis had found). Move the dispenser to a fast self-serve location so commuters can grab it quickly before the drive. Both improvements come directly from the job — not from asking customers what they wanted in a milkshake, but from understanding what they were hiring it to do.
What made this possible
The investigation focused on the circumstance of hiring, not the product’s features or the customer’s profile. The counterintuitive job — make a boring commute bearable, one-handed, until lunch — was invisible until someone studied what was actually happening at 7am in the drive-through. The job cannot be guessed from the product. It can only be found in the situation.
Frameworks
Where Jobs To Be Done shows up.
JTBD is a demand-understanding lens, so it maps to the moments in each framework where the task is deciding what to build and for what underlying need — before the team commits to a direction.
Related methods
What to combine with JTBD.
Sources & further reading
The work behind this method.
Competing Against Luck
Clayton Christensen, Taddy Hall, Karen Dillon, and David Duncan (2016)
The definitive book on Jobs To Be Done and the source of the milkshake study. Christensen lays out the full theory of why people hire products and what drives switching, with the milkshake research as its most famous illustration. The book makes the counterintuitive argument — that the job, not the product or the customer, is the unit of innovation — in its most developed form.
Jobs to Be Done
Anthony Ulwick (2016)
Ulwick's Outcome-Driven Innovation variant of the theory, which operationalizes JTBD with a more quantitative methodology for identifying and prioritizing jobs. Where Christensen's version is more conceptual, Ulwick's provides structured techniques for mapping jobs to outcomes and scoring them against importance and satisfaction. A complement to Competing Against Luck for teams who want more process.
When Coffee and Kale Compete
Alan Klement (2016)
Klement's demand-side perspective on JTBD focuses on the emotional and social dimensions of the job — the parts that product teams most consistently miss. His framing of the job as a system of progress, involving not just the functional task but the emotional and social progress the person is trying to make, is the clearest treatment of why the method reaches deeper than features or demographics.