Jobs to Be Done: The Framework Explained for Startups, With Examples and Interview Questions
What a job to be done is, how to write one, the questions that uncover it, how it differs from a persona, and how to use it when you have no customers yet.
By Suren, Founder & Growth Strategist at Gufy · Published
A job to be done is the progress a person is trying to make in a particular situation. People do not buy a product for its own sake. They hire it to get a job done, and they will swap it for anything that does the job better.
For a startup this is the most practical idea in marketing, because it tells you what to build first, what to say, and who you are really competing with. This guide explains the framework, shows how to write a job, gives the interview questions that uncover one, and covers how to use it before you have customers.
Where the idea comes from
The framework was developed by several people. Clayton Christensen made it widely known and set it out in Competing Against Luck. Tony Ulwick built a detailed method around it called Outcome-Driven Innovation. Bob Moesta developed the interview approach and the "forces" described below.
The older version of the idea is a line usually credited to Theodore Levitt: people do not want a quarter-inch drill, they want a quarter-inch hole.
The milkshake
Christensen's best-known example involves a fast-food chain that wanted to sell more milkshakes. It asked customers what would improve the product and made the changes. Sales barely moved.
Then the researchers watched who was buying. A large share of milkshakes were sold in the morning, to people on their own, who took them to the car. Asked why, they explained that they had a long, dull commute, wanted something to do with one hand, and needed to stay full until lunch. A banana was gone in a minute. A bagel made a mess. A thick milkshake lasted the whole drive.
Two lessons come out of it, and both apply directly to a startup:
- The job was not "a tastier milkshake". It was "make my commute less boring and keep me going until lunch". Improving the product on the wrong measure did nothing.
- The competition was not other milkshakes. It was everything else that could do the same job in that situation.
The three parts of a job
A job usually has three layers. All three affect the decision to buy.
| Layer | What it is | Example: small-business owner choosing accounting software |
|---|---|---|
| Functional | The practical task | Send invoices and know who has not paid |
| Emotional | How they want to feel | Stop worrying at night about cash |
| Social | How they want to be seen | Look professional to clients and to the accountant |
Founders tend to write the functional layer and stop. The emotional layer is often what gets someone to act.
How to write a job
A simple format, sometimes called a job story:
When [situation], I want to [motivation], so I can [outcome].
Examples:
- When a customer asks for a quote, I want to send an accurate price the same day, so I can win the work before a competitor replies.
- When I finish a week of client work, I want the admin to be done already, so I can have my weekend back.
- When a new matter comes in, I want to know quickly whether it is worth taking, so I can spend my hours on work that pays.
Three checks:
- The product is not in the sentence. A good job would be true if your product did not exist.
- The situation is specific. "When I am running a business" is not a situation. "When a customer asks for a quote" is.
- The person would say it out loud. If it sounds like a strategy document, rewrite it in their words.
A quick test: which basic need does it serve?
A rule of thumb we use on founder calls. Most products people pay for, and keep using, help with one of a small number of things:
- making money
- saving money
- saving time
- staying healthy
- connecting with other people
Look at the apps you open every day and each will fit one of these. If you cannot say which one your product serves, for a specific person, you have not found the job yet.
The same test explains why one product needs different messages for different customers. An owner working seven days a week wants time. A business paying for five separate tools wants a smaller bill. Same product, different job.
People already have a way of doing it
This is the point founders most often miss. Your customer is getting the job done today, without you. They have a spreadsheet, an assistant, a general-purpose AI tool, a habit, or they have decided to live with the problem.
So the question is never "would this be useful?" Almost everything would be useful. The questions are:
- How are they doing it now, and what does that cost them?
- How much better is your way? A small improvement does not overcome the effort of switching. Our positioning guide has an exercise for scoring this.
- Is your product a must-have or a nice-to-have for this person?
A founder building an app for new parents had confirmed, with a survey, that parents wanted support and information in one place. What the survey had not asked was what they used today. The answers were search, AI assistants, social groups and friends, all free. The job was real. The open question was whether bringing those things together was worth a monthly fee, and the survey had not tested that.
People already have set lives. It is tempting to build a product and then look for a place in the customer's day where it might fit. It works the other way round: learn the shape of their day, then cut the product to fit it.
The four forces
Bob Moesta's model explains why people switch, or do not, even when a better option exists. Four forces act on every decision.
| Force | Direction | What it sounds like |
|---|---|---|
| Push of the current situation | Towards switching | "I cannot keep doing quotes at midnight." |
| Pull of the new solution | Towards switching | "This would send it in twenty minutes." |
| Anxiety about the new solution | Against switching | "What if moving my data over is a nightmare?" |
| Habit of the present | Against switching | "My spreadsheet is ugly but I know where everything is." |
A sale happens when push and pull together outweigh anxiety and habit. Most marketing only works on pull, by describing the product. Often the faster gains come from reducing anxiety: a free trial long enough to see the value, help with setup, or a plain statement of how long a step takes. Our guide to the aha moment covers that part.
Anxiety can also be attached to a single word. In a profession that has seen public failures of AI tools, leading with "AI" raises a barrier before you have said what changes for the buyer. Lead with the progress they want.
Jobs to be done and personas
| Persona | Job to be done | |
|---|---|---|
| Describes | Who the customer is | What they are trying to get done, and when |
| Built from | Demographics, role, habits | Situations, frustrations, desired outcomes |
| Good for | Targeting ads, choosing channels, tone | Deciding what to build and what to say |
| Weakness | Two people in one persona can want different things | Does not tell you where to find people |
Use both. The ideal customer profile tells you who to go after. The job tells you what to say when you reach them.
The user's job and the buyer's job
In business software the person who uses the product and the person who pays for it often have different jobs. A founder with an AI tool for lawyers had built his message around a technical strength of the product. The lawyer using it might care about that. The principal of the firm, who would approve the purchase, had a different job: more billable work from the same team, without more risk.
Write a job for each. Your product has to do the user's job. Your sales message has to speak to the buyer's.
How to find the job
If you have customers
Interview people who bought recently, and people who stopped. Ask about what happened, not what they think.
- What was going on when you first started looking for something like this?
- What were you using before? What was wrong with it?
- What else did you consider, including doing nothing?
- What nearly stopped you from signing up?
- What can you do now that you could not do before?
- If we switched it off tomorrow, what would you do instead?
Ask about a specific past event. "Tell me about the last time you had to do this" gets you facts. "Would you use a tool that…" gets you politeness.
If you have no customers yet
- Write your guesses down as hypotheses. For example: "Lawyers want an AI tool they can trust not to invent things", "Lawyers want to spend less time on document analysis", "Firm principals want more billable work per person". Each is a guess until tested.
- Talk to five people in one specific group who do not know you. People you already deal with have reasons to be kind. A message on LinkedIn, an introduction or an industry event will get you five conversations.
- Ask how they do it today, using the questions above. Do not pitch. If you describe your solution first, they will agree with you, and you will learn nothing.
- Record the conversations if they agree. You will hear what confirms your idea. Someone else listening later will hear what you missed.
- Count the pattern. If four out of five say the task you want to remove is the part of the job they value, your market is the fifth person. That may still be a business. It is a different one from the business you planned.
- Ask for a commitment. Until a stranger pays, the job is a hypothesis. Strangers paying is also the first sign of product-market fit.
The SaaS idea validation playbook covers the rest of the process.
What to do with the job once you have it
- Build for it first. The first version of the product should do the job and little else. See What Is a Minimum Viable Product?
- Open your message with it. The first job of your headline is to make the right person stop and think "that is me". Features come after you have their attention.
- Redraw your competition. List everything that does the job today, including free tools and doing nothing. That list, not your category, sets what you can charge.
- Design onboarding around it. The moment a new user first gets the job done is the moment they have a reason to stay.
- Use it to choose a segment. The group that feels the push most strongly is the one to start with.
Common mistakes
- Writing the job around your product. "I want an all-in-one dashboard" is a solution, not a job.
- Stopping at the functional layer. Ask how they want to feel when it is done.
- Asking about the future. People are poor at predicting what they will do and good at describing what they did.
- Interviewing people who know you.
- Ignoring the free alternatives.
- Treating a survey as proof. It can confirm the problem. It rarely confirms that anyone will pay.
Common questions
What is a job to be done?
A job to be done is the progress a person is trying to make in a particular situation. The idea is that people do not buy a product for its own sake. They 'hire' it to get something done, and they will replace it with anything that does that job better, including things that look nothing like your product.
What is an example of a job to be done?
The best-known example is Clayton Christensen's milkshake. A fast-food chain found that many milkshakes were bought in the morning by commuters who wanted something to make a long drive less dull and keep them full until lunch. The job was the commute, so the real competition was a banana, a bagel or a doughnut.
How do you write a job to be done statement?
Use the form: When [situation], I want to [motivation], so I can [outcome]. For example: 'When a customer asks for a quote, I want to send an accurate price the same day, so I can win the work before a competitor replies.' Keep the product out of the sentence. A good statement would still be true if your product did not exist.
What is the difference between jobs to be done and personas?
A persona describes who the customer is: role, age, habits. A job describes what they are trying to get done and in what situation. Two very different people can have the same job, and the same person has different jobs at different moments. Personas help you find and address people. Jobs tell you what to build and what to say.
How many interviews do you need to find the job to be done?
Five conversations with people in one specific group, who do not know you, is enough to see whether the same situation and the same frustration keep coming up. If all five describe different problems, narrow the group and try again before drawing any conclusion.
Where Gufy fits
Gufy works with SaaS and mobile-app startups on go-to-market strategy: turning customer conversations into a clear job, a message built on it, and tests with real buyers through landing pages and paid campaigns. A growth workshop is often where the job gets written down for the first time.
If you have built the product and are not sure what it is really being hired to do, bring your interview notes or recordings.