Skip to main content

UX and research · Beginner · 14 min

Write job stories with jobs to be done

By Padding · Updated 18 Nov 2026
Any notes app or document, plus a spreadsheet · November 2026

You will have a checked set of job stories for booking a clinic appointment, written from interview notes, set side by side with the user stories they replace, and a note on where they agree and clash with your personas.

On this page

A user story says who wants something: "As a patient, I want to book online, so that I save time." It reads well in a backlog, but the "as a patient" part does very little work. Every patient is a patient. The story does not say what was happening when they reached for the app, and that moment is usually what decides whether a design works.

A job story starts from the moment instead of the role:

When [situation], I want to [motivation], so I can [outcome].

The idea behind it is jobs to be done: people do not want a product, they want to make some progress in a particular circumstance, and they "hire" whatever helps them make it. A job story is a small, written record of one of those circumstances. It gives designers and engineers something to test a screen against: does this help the person in that situation make that progress?

In this tutorial you turn raw interview notes into job stories, check each one, rewrite the user stories for the same feature and compare the two, and then hold your stories up against a persona to see where they help each other and where they pull apart.

The running example is Larkfield Clinic, a fictional community clinic building an appointment booking app: choose a service, pick a time, enter your details, confirm, and get a reminder. Use your own notes if you have them; the steps work the same way.

Step 1 · Mark the moments in your interview notes

Job stories come from things people did, not things they said they would like. Go through your notes and mark every passage where someone describes a specific occasion: a time, a place, something that had just happened, and what they did next.

If you have no notes of your own, use these extracts from three Larkfield Clinic interviews:

P1, shift worker, 34. "I only find out my rota on Thursday for the week after. So I ring on Friday morning and they say there's nothing until the week after that. Last time I just went to the walk-in place in town." >P2, parent of two, 41. "My son had an ear infection and I was trying to book while he was crying in the back of the car. I had the phone in one hand. I gave up and called." >P3, retired, 68. "I booked my blood test in the app, but I wasn't sure it had gone through. No email came. So I phoned to check, and the receptionist said yes, it's there." >P1 again. "I'd booked one for a Tuesday, then got moved to a late shift. I didn't know how to cancel without ringing, so I just didn't go. I felt bad about it." >P2 again. "When I booked his check-up online last month, the screen just went back to the start. I rang the next morning to make sure it was actually booked."

Copy each marked passage into its own row of a spreadsheet, with columns for Participant, Quote and, for now, an empty Situation column. Leave out opinions with no occasion attached ("apps are fine I suppose"). They are not wrong, but they do not describe a moment you can design for.

Check it worked

Your sheet has one row per occasion, and every quote in it mentions something that happened (a shift change, a sick child, a missing email). If a row only holds a preference or a general complaint, delete it or find the occasion behind it in the full notes.

Step 2 · Write the situation in plain words

For each row, fill the Situation column with what was true at that moment. Include the trigger (what had just happened) and any constraint that shaped what they could do (time, place, hands free or not, what they knew).

From the sample notes:

  • P1: "my shifts for next week have just been published and I need an appointment that fits around them"
  • P2: "my child is unwell and I am trying to book on my phone with one hand, away from home"
  • P3: "I have just booked in the app and have not received any sign that it worked"
  • P1 again: "my shift has changed and I can no longer make an appointment I booked"
  • P2 again: "I have just booked online and the app gave me nothing to show it worked"

Keep the person's role out of it. "As a shift worker" is a label; "my shifts for next week have just been published" is a situation. Two very different people can share a situation, and one person moves through several in a week.

Check it worked

Read each situation aloud and ask: could I set up this moment in a usability test? "My child is unwell and I am booking one-handed, away from home" can be acted out. "I am a busy parent" cannot. Every situation in your sheet passes that test.

Step 3 · Add the motivation and the outcome

Add two more columns: Motivation (what they want to do in that moment) and Outcome (the progress it gets them, in their terms). Then join the three parts into a sentence in a Job story column.

For the Larkfield Clinic notes:

  1. When my shifts for next week have just been published, I want to see which slots fall outside my working hours, so I can book before the good times go.
  2. When my child is unwell and I am booking on my phone away from home, I want to get to an available slot with as few taps as possible, so I can stop worrying about whether they will be seen today.
  3. When I have just booked an appointment, I want to see clear proof that it is confirmed, so I can stop thinking about it without phoning to check.
  4. When my plans change after I have booked, I want to move or cancel the slot from where I booked it, so I can free it up without a phone call and without feeling I have let the clinic down.

Two rules keep these honest. The motivation names what the person is trying to do, not a feature ("see which slots fall outside my working hours", not "filter by time"). The outcome names a change in their life or state of mind, not a change on the screen ("stop thinking about it", not "see a confirmation page").

Check it worked

Cover the motivation of each story and read the situation and outcome alone. If you could think of two or three different designs that would get there, the story leaves room for design. If only one design fits, a feature has slipped into the motivation; rewrite it.

Step 4 · Test each story against the notes

A job story is only as good as the evidence behind it. Add a column called Evidence and, for each story, write how many participants described that situation, and which ones.

Then check each story for these common faults:

  • Invented situation. Nobody in the notes was in it; you imagined it. Either remove the story or mark it "assumption" and plan to ask about it in the next round of interviews.
  • Two jobs in one. The story has "and" joining two motivations. Split it.
  • Outcome is the motivation repeated. "I want to book quickly, so I can book quickly." Ask "and then what?" until you reach a real change for the person.
  • Emotional side missing. People often want to feel or avoid feeling something (P1 "felt bad" about not turning up). If the notes show it, put it in the outcome.

Story 4 above has two motivations (move or cancel) but one situation and one outcome, so it is fine as one story.

In the sample notes, only story 3 has two participants behind it: P3 phoned to check a booking, and P2 did the same after a check-up booking. Stories 1, 2 and 4 each rest on one person (P1, P2 and P1), so mark all three as weak. That is normal after three interviews; "weak" means "ask about this next round", not "delete it".

Check it worked

Every row has an Evidence entry with at least one participant code, or is marked "assumption". No story has two separate motivations joined by "and" that would lead to different screens.

Step 5 · Compare with the user stories for the same feature

Find or write the user stories a backlog would usually hold for the same feature. For Larkfield Clinic, a typical set looks like this:

  • As a patient, I want to book an appointment online, so that I do not have to phone.
  • As a patient, I want to receive a confirmation, so that I know my booking is made.
  • As a patient, I want to cancel my appointment, so that someone else can have it.

Put them in a second list next to your job stories and compare them on four questions. Write a short answer for each:

  1. What would a designer build from it? The first user story gives a booking form. Job story 1 gives a way to see times against a known weekly pattern; job story 2 gives a one-handed, low-effort path. These are different screens.
  2. What does it leave out? None of the user stories mentions the phone calls P2 and P3 made after booking, which is the real cost the clinic wants to remove.
  3. Whose reason is in the "so that"? The third user story's reason ("someone else can have it") is the clinic's, not the patient's. Job story 4 keeps the patient's own reason, and the clinic's benefit follows from it.
  4. Can you test it? A job story gives you a test scenario directly. A user story gives you a role and a feature, and you still have to invent the scenario.

User stories still have a place. They are a convenient unit for planning and tracking work, and many teams keep them. A common approach is to keep the job story as the reason and write the delivery tickets beneath it.

Check it worked

Your comparison has an answer for each of the four questions for at least two pairs of stories, and you can point to one thing a job story revealed that none of the user stories mentioned.

Step 6 · Hold the stories up against a persona

Personas and job stories describe different things. A persona describes a person: their background, habits, limits and goals across many situations. A job story describes a moment, whoever is in it. Used together, they cover each other's gaps, but they can also contradict each other.

If your team has a persona, put it next to your stories. For Larkfield Clinic, imagine a persona called "Busy parent Ada" whose summary says she wants "fast, simple booking".

Go through your stories and sort each into one of three groups:

  • Persona adds detail. The persona tells you something that changes how you meet the job. Ada uses an older phone with a small screen, which matters for job story 2.
  • Story goes beyond the persona. The job applies to people the persona does not describe. Job story 3 (proof of booking) came from P2, a parent, but also from P3, a retired person. If the team only designs "for Ada", the older patients who share this job risk being forgotten.
  • They clash. The persona says one thing and the evidence says another. Ada's "fast, simple booking" suggests speed is everything, but job story 1 (from a shift worker) wants more information on the screen, not less. Write down which one has the stronger evidence.

Where they clash, trust the one built on more and newer research. A persona written from a workshop with no interviews behind it should give way to job stories with participant evidence.

Be honest about the example clash, though: job story 1 rests on P1 alone. One participant is enough to question the persona, not to overturn it, so record the decision as provisional and take the question (does showing more slot detail help or slow people down?) into the next round of interviews.

Check it worked

Each story has one of the three labels. For every clash, you have a sentence saying which source you are trusting and why, with participant codes.

Step 7 · Finish the set for booking an appointment

Arrange your checked stories in the order a patient meets them, and map each to the app screens it affects. For Larkfield Clinic:

  1. New shifts published, see slots outside working hours. Screens: Choose a service, Pick a time. Evidence: P1 (weak).
  2. Child unwell, booking one-handed. Screens: Home, Pick a time, Your details. Evidence: P2 (weak).
  3. Just booked, want proof. Screens: Confirm, Booked, the reminder. Evidence: P2, P3.
  4. Plans changed, move or cancel. Screens: Booked, the reminder. Evidence: P1 (weak).

Add any screen with no job story pointing at it to a list called Unsupported screens. That is not proof the screen is unnecessary, but it is a question for the team: who needs it, and in what moment?

Share the list with whoever writes tickets or designs the screens, and ask them to name the job story each piece of work serves.

Check it worked

Your final document has a numbered set of job stories, each with evidence, a list of screens, and a persona label from the previous step, plus a short list of unsupported screens (empty is a valid answer, but you looked).

Next steps

Need this done for you?

Padding does this work for growing companies, from the first diagnosis to a system that keeps working.

Digital Transformation

Get new tutorials

New tutorials, about twice a month.