Jobs To Be Done (JTBD) in growth strategy: positioning on what customers are really trying to do

People buy progress, not products

Most businesses describe themselves by what they sell. Features, integrations, price plans. The customer, meanwhile, is not shopping for features. They are stuck somewhere in their week and want to be somewhere better. The product is only the vehicle.

Jobs To Be Done (JTBD) is a way of looking at customers that starts from that stuck moment. The Christensen Institute, which carries on Clayton Christensen's work, describes a job as the progress people are trying to make as they strive towards a goal "within particular circumstances".¹ It adds that every job has functional, social and emotional sides.¹ Christensen and his co authors set the idea out for managers in the Harvard Business Review in 2016, under the title "Know Your Customers' Jobs to Be Done", arguing that innovation stops being a game of chance once you understand why customers choose what they choose.²

The word that does the work is circumstance. Two people with the same age, income and job title can buy the same product for completely different reasons, because they are in different situations. One person can buy it for one reason in April and another in October. That is why a job tells you more about what to say, and when, than a persona does. I wrote about that wider point in segmentation is not demographics.

Throughout this article, one invented business carries the examples. Every name and number in it is made up.

Worked example: what Leeside Desk thinks it sells

Leeside Desk is a small software company in Cork. Róisín runs it with one developer. It sells scheduling and invoicing software to sole trader plumbers and electricians at €29 a month, from leesidedesk.example.

The homepage headline reads "All in one job management for trades". The feature list covers a calendar, quotes, invoices, a customer database and reports. Róisín assumes her customers want a tidy system.

Now look at it as a job. A plumber finishes a boiler repair at 4.30pm on a Friday. The customer says "send me the invoice". The plumber means to do it that evening, forgets, and is paid three weeks later after two awkward phone calls. What this plumber wants is not a tidy system. It is to be paid for Friday's work without chasing anyone.

• Functional side: get the invoice out before leaving the driveway.
• Emotional side: stop the Sunday evening dread about who still owes money.
• Social side: look like a proper business to the customer, not someone writing figures on the back of a receipt.

Same product, a very different first sentence on the homepage.

A practitioner framework, not a measured law

It is worth being straight about what JTBD is: ideas and interviewing methods built by practitioners from case work. It is not a result measured in controlled studies, and nobody can tell you that applying it lifts conversion by a given percentage. When an article claims that it does, ask where the number came from.

There are three main schools, and they share the word "job" while doing quite different things.

1. Christensen's theory of jobs. The idea that customers "hire" a product to get a job done, and that the circumstance, not the customer's profile, explains the choice.²¹ It is a way of thinking, taught through cases.
2. Tony Ulwick's Outcome Driven Innovation (ODI). A structured process that breaks a job into steps, collects the measures customers use to judge success, and surveys customers to score which needs are underserved.³ By Strategyn's own account, Ulwick conceived the approach in 1990 and named the process ODI in 1999.³ Of the three, it is the most quantitative.
3. Bob Moesta's switch interviews. A way of interviewing people who recently switched from one product to another, reconstructing the timeline of their decision and the forces that pushed and held them.⁴⁵ Moesta and Chris Spiek developed the four forces model as part of that method.⁴

The success figures you see quoted come from the people selling the methods. Strategyn reports that 18 of 21 product launches that used ODI were rated successful, about 86%.⁶ Those ratings were given by the companies that sponsored the projects.⁶ That is useful as a practitioner's track record. It is not independent evidence, and 21 launches is a small sample.

Worked example: the milkshake, as Christensen told it

The most famous JTBD story has grown in the retelling, so here is only what the published sources say.

In 2011, Harvard Business School's Working Knowledge reported a teaching note by Christensen. A researcher working with him studied a fast food chain that wanted to sell more milkshakes.⁷ He found that 40% of milkshakes were bought first thing in the morning, by commuters who took them to go.⁷ They faced a long, boring drive, wanted something to keep them going until mid morning, and had at most one free hand.⁷ The milkshake was competing less with other milkshakes and more with a bagel or a doughnut.⁷ The Christensen Institute adds that later in the day the same product did a different job, a treat parents bought for their children.¹

What the sources do not report is a sales result. Versions online that quote a precise sales lift, or name the chain, go beyond what Christensen published, so I leave them out. The story still teaches the point: the same product, two circumstances, two jobs, two different competitors.

Writing the job down in one sentence

A job that lives only in your head drifts. I write each one down in a fixed shape so that it stays specific:

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

This shape is my working habit, not a rule from any of the three schools. It forces the circumstance, the motivation and the outcome onto the page, and keeps the product out of the sentence.

Three tests I apply to every statement:

• Could a customer have said it, in those words, before your product existed?
• Does it name a circumstance you could find on a calendar or a map?
• Would a competitor that looks nothing like you also be able to do this job?

Worked example: three job statements for Leeside Desk

Róisín's first attempt: "When I need job management software, I want an all in one tool, so I can be organised." It fails all three tests. It names a category, there is no circumstance, and the outcome is vague.

Her second attempt, after listening to customers (the next section shows how): "When I finish a job and the customer is standing there, I want to send the invoice before I leave, so I can be paid this week without chasing."

That passes. A customer could have said it in 2010. The circumstance is a real place and time. And the competitors include a paper invoice book, a banking app with payment links, and a spouse who does the paperwork of a Sunday. None of those is "job management software".

She writes two more:

• "When a job overruns, I want to move the rest of my week in a couple of minutes, so I can text the next customers before they wonder where I am."
• "When the year end comes, I want my figures ready for the accountant, so I can stop paying them to sort out a shoebox."

Three jobs, one product. Each one will need its own message.

Finding the job: talk to people who just switched

Jobs are not found by asking customers what they want. People are good at telling you what happened and poor at predicting what they would do. So the Moesta school interviews people about a decision they have already made.

The guidance on the method is specific. Interview people who switched in the last 30 to 90 days, because beyond about 90 days the details are gone.⁵ Plan for 45 to 90 minutes.⁵ Move forward through concrete events, not opinions, with questions like "when did you first start thinking about this?" and "and then what happened?"⁵ Include people who almost switched but did not, and avoid your happiest long standing customers, whose story is too settled to be useful.⁵ Patterns often show up within five to eight interviews for a given job.⁵

In my experience the hardest part is resisting the urge to pitch, or to correct people when they describe your product wrongly. Their wrong description is data.

Ask permission before you record anything, say what the recording is for, and delete it when the notes are written.

Worked example: Róisín's six interviews

Róisín emails 15 people who started paying in the last two months and six who tried and left. Seven agree to talk, five payers and two leavers, for a €20 voucher and 45 minutes.

Her opening question is the same each time: "Take me back to the day you first thought, I need something other than what I have now. Where were you?"

Here is a condensed version of one answer, from a heating engineer in Carrigaline, Co. Cork:

1. "It was February. I had four invoices out over 30 days and my van needed tyres."
2. "A pal showed me his phone in the queue at the builders' merchants. Job done, invoice gone, paid by card on the spot."
3. "I looked at three apps that night. Two wanted me to set up a whole company profile first."
4. "I almost didn't bother. I've used the same carbon copy book for 11 years."
5. "What did it was when a customer paid me on the doorstep using the link. I signed up for the paid plan that evening."

Not one of the seven named reports, the customer database or the calendar as the reason they started. Five described a moment of not being paid. Róisín's feature list would never have shown her that.

Every switch follows a timeline

Moesta and Spiek describe the decision to switch as a sequence of stages.⁸ It starts with a first thought, usually earlier than people remember. Then comes passive looking, when the person notices things but is not shopping. A first event turns "I should look into this" into "I need to do something". Active looking follows, and the shopping itself sharpens what they want. A second event, often a deadline or a nudge, moves them to deciding. After the purchase comes consumption, first in the mind and then in reality, and the gap between the two decides whether they are satisfied.⁸

Read the diagram from the top: each box is a stage in one decision, in order. The takeaway is that a customer is reachable at every stage, but the message that works differs. A passive looker needs a gentle signal. An active looker needs a clear comparison. Someone at the second event needs the last doubt removed.

Worked example: the heating engineer's timeline

Róisín maps the Carrigaline interview onto the stages:

1. First thought: February, four invoices over 30 days.
2. Passive looking: seeing a pal's phone in the queue at the builders' merchants.
3. First event: the tyres. Money owed now has a cost.
4. Active looking: three apps in one evening, two rejected for a long setup.
5. Second event: a customer pays on the doorstep with a link during the trial.
6. Deciding: the paid plan the same evening.
7. Using it: the first month, where the invoice either goes out from the driveway or it does not.

Two lessons fall out of it. Setup must be short enough to finish on a phone in an evening, where active looking happens. And the trial must reach a paid invoice quickly, because that was the second event.

Four forces decide whether anyone switches

Moesta and Spiek describe every switch as a tug of war between four forces.⁴ Two pull the customer towards the new product:

• The push of the situation. The problem with how things are now.⁴
• The pull of the new solution. The better future the customer imagines, which is not the same as your feature list.⁴

Two hold the customer back:

• Anxiety about the new solution. Fear of the unknown: will it work, will I look foolish, what about my data.⁴
• The habit of the present. The comfort of what they already know, even when it is not great.⁴

A customer switches only when the push and the pull together outweigh the anxiety and the habit.⁴ Moesta's own consultancy frames the forces in the same way: pushes come from the struggling moment, pulls describe what progress looks like, anxieties concern the new future, and habit is the old thing they must give up.⁹

This is not a sum you can calculate, because nobody measures anxiety in the same unit as habit. It is a checklist for interviews and copy. Most marketing works only on the pull. In my experience the better value win is often to reduce the anxiety.

Follow the diagram down: three boxes for the forces, a diamond for the comparison the customer makes, mostly without noticing, and two outcomes. The takeaway sits in the last box: when people stay where they are, adding more pull is rarely the fix.

Worked example: sorting Róisín's notes into forces

Róisín goes back through her seven interviews and files every remark under one force.

• Push: unpaid invoices, chasing by phone, the year end shoebox. Five of seven.
• Pull: a friend's phone, being paid on the doorstep. Four of seven.
• Anxiety: "will it work with my accountant's software", "I'm not great with apps", "what if I send the wrong amount". Six of seven, including both people who left.
• Habit: the carbon copy book, a spreadsheet a partner keeps. Four of seven.

The pattern is plain. The push is strong, the pull is grand, and anxiety is what stopped the two who left. So Róisín's next work is not a new feature. It is a short line under the sign up button saying that invoices can be checked before they are sent, a page showing which accounting packages it exports to, and a two minute video of a real invoice going out from a van.

Scoring which needs are underserved

Interviews tell you what the jobs are, not how many customers share each need or which is worst served. That is what Ulwick's side of JTBD is for.

ODI breaks the core job into steps. Strategyn's generic job map has eight: define, locate, prepare, confirm, execute, monitor, modify and conclude.³ For each step it collects desired outcomes, written in a fixed shape: a direction such as minimise, a metric, the object being controlled, and a clarifier for the context.¹⁰ Ulwick's own example, for the job of listening to music, is "minimize the time it takes to get the songs in the desired order for listening".¹⁰ Strategyn says it collects 5 to 10 outcomes per step and 100 or more for a core job.¹⁰

Customers then rate each outcome twice in a survey: how important it is, and how satisfied they are with how they do it today. Strategyn turns each into the share of respondents who give one of the top two answers, "very" or "extremely".¹¹ It combines the two in what it calls the opportunity algorithm.¹²

$$O = I + \max(0,\ I - S)$$

Here $O$ is the opportunity score of one outcome. $I$ is its importance, the share of respondents rating it very or extremely important, written on a scale of 0 to 10 (so 84% becomes $I = 8.4$). $S$ is its satisfaction, the share rating themselves very or extremely satisfied, on the same scale. $\max(0,\ I - S)$ takes the gap between importance and satisfaction, and sets it to zero when satisfaction is higher, so an overserved need is never pushed below its importance. Converting a share to a score is plain percentage work, explained in Foundations: percentages.

Importance counts twice. A need that matters and is badly met scores highest. A badly met need nobody cares about stays low.

Worked example: Róisín's survey

Róisín turns her interviews into five outcome statements and sends a short survey to all customers and recent trial users. 160 people answer. For each statement she counts the share who said very or extremely important, and very or extremely satisfied with how they handle it today.

Outcome	Important	Satisfied
A. Minimise the time it takes to send an invoice after finishing a job	84%	31%
B. Minimise the likelihood of forgetting to chase an unpaid invoice	78%	22%
C. Minimise the time it takes to rearrange the week when a job overruns	66%	40%
D. Minimise the effort of preparing figures for the accountant at year end	58%	45%
E. Minimise the time it takes to produce a branded quote	41%	63%

Outcome A, step by step. Importance 84% becomes 8.4, satisfaction 31% becomes 3.1.

$$O_A = 8.4 + \max(0,\ 8.4 - 3.1) = 8.4 + 5.3 = 13.7$$

The gap $I - S = 5.3$ is positive, so it counts in full, and $O_A = 13.7$.

Outcome B: importance 7.8, satisfaction 2.2, gap 5.6.

$$O_B = 7.8 + \max(0,\ 7.8 - 2.2) = 7.8 + 5.6 = 13.4$$

Again the gap is positive, so $O_B = 13.4$.

Outcome C: 6.6 plus a gap of 2.6 gives 9.2. Outcome D: 5.8 plus a gap of 1.3 gives 7.1. Outcome E is the interesting one. Importance is 4.1 and satisfaction 6.3, so the gap is negative and the max sets it to zero.

$$O_E = 4.1 + \max(0,\ 4.1 - 6.3) = 4.1 + 0 = 4.1$$

Here $I - S = -2.2$, the max turns it into zero, and $O_E$ equals its importance alone.

Along the bottom are the five outcomes. Up the side is the opportunity score, importance plus the unmet gap, on a scale that runs from 0 to 20. The takeaway: the two outcomes about getting paid stand well clear of the rest, and the branded quote, which Leeside Desk's homepage currently shows first, is the least underserved need of the five.

Worked example: is A really ahead of B?

A scores 13.7 and B 13.4. It is tempting to rank them. With 160 respondents, the shares themselves carry sampling error, so check before you do.

The standard error of a share tells you how much it would wobble from one sample of 160 to the next.

$$SE = \sqrt{\frac{p(1-p)}{n}}$$

$SE$ is the standard error, $p$ is the share as a decimal (0.84 for 84%), and $n$ is the number of respondents. The idea is explained in Foundations: sampling, and the 95% range built from it in Foundations: confidence intervals.

For the importance of A, with $p = 0.84$ and $n = 160$:

$$SE_A = \sqrt{\frac{0.84 \times 0.16}{160}} = \sqrt{0.00084} = 0.029$$

A 95% range is the share plus or minus 1.96 standard errors: $0.84 \pm 1.96 \times 0.029$, which is 78.3% to 89.7%. For B, $SE_B = \sqrt{0.78 \times 0.22 / 160} = 0.033$, giving 71.6% to 84.4%.

The two ranges overlap heavily. A and B are effectively tied. That is fine, because they are two sides of the same job: getting paid for work already done. Róisín does not need to choose between them. She needs to lead with that job.

Turning the job into positioning

Once the job is clear, positioning is a matter of saying it back to the customer in their own words, in the place where they are looking. I work through the same four questions each time:

1. What is the struggling moment, in the customer's words?
2. What are they using today instead, and why does it fall short?
3. What does progress look like to them?
4. What anxiety will stop them, and what removes it?

The answers feed the homepage, the ads, the trial and the order of features. For more on the copy, the StoryBrand framework and the message, market and model fit article both pick up from here.

Worked example: Leeside Desk's new homepage and ads

The homepage before: "All in one job management for trades".

The homepage after: "Send the invoice before you leave the driveway. Get paid this week, without chasing." Underneath, one line for the anxiety: "Check every invoice before it goes. Exports to the accounting software your accountant already uses."

The trial changes too. Instead of a company profile, a calendar import and a tour of reports, it asks for a name and a logo and goes straight to "send your first invoice". I go deeper on that ordering in onboarding and habit formation.

The ads split by trigger, not by trade:

• For the push of unpaid work: "Four invoices over 30 days? Send them from the van and get paid by card."
• For the habit of the carbon copy book: "Keep the way you work. Lose the book."

Each ad leads to a page that opens with its own job. Whether any of this works is an empirical question, not a JTBD one. Róisín runs the homepage change as an A/B test, sized in advance, which I explain in how much data you actually need. If her traffic is too small for a fair test, she calls the result an honest before and after, nothing more.

The same customer has different jobs

The circumstance changes, and the job changes with it. I find this most useful in eCommerce, where the same shopper arrives with a different job in a different month.

So one homepage and one ad cannot serve every job. You do not need a version for every person, only for every circumstance that matters to your numbers.

Worked example: a Galway gift shop across the year

Quay Lane Gifts is an invented online gift shop in Galway, at quaylanegifts.example. The owner, Aoife, interviews eight recent buyers and finds three jobs that recur.

• Early December: "When I have six people to buy for and no ideas, I want a short list I can trust, so I can be done in one evening." Message: gift guides by recipient, delivery dates up front.
• The week before a birthday: "When I have left it late, I want something that looks chosen rather than grabbed, so I don't look like I forgot." Message: next day delivery, a handwritten card option.
• A thank you after a favour: "When a neighbour has helped me out, I want a small gift under €25 that says thanks properly, so I can return the favour without making it awkward." Message: a thank you collection, gift wrapped by default.

Same shop, three circumstances, three landing pages with matching ads and emails. In December the gift guide leads the homepage. The rest of the year, the late birthday job does.

Where JTBD goes wrong

The method is simple to describe and easy to bend. The mistakes I see most:

Interviewing only happy, long standing customers. Their memory of the switch is gone, and they have no reason to mention anxiety. Talk to recent switchers and to people who nearly switched.⁵

Turning jobs back into personas. "Busy Brian, 42, plumber" is a persona with a job glued on. The circumstance is the point.

Treating seven interviews as a measurement. Interviews find jobs, not how common each one is. For that, survey, and check the sampling error.

Quoting the method's success rates as proof. The published figures come from the people selling the methods. JTBD earns its keep in your own tests.

What to take away

Customers do not buy your product. They pull it into their life to make progress in a particular circumstance.¹ Write that progress down in one sentence, find it by talking to people who recently switched, map the timeline and the forces, and, if you need to rank needs, score them with care. Then say the job back to the customer, where they are looking, in their words.

None of it needs a big research budget. Róisín's version took seven conversations, a 160 person survey and a fortnight, and the patience to listen before writing a headline.

Footnotes

1. Clayton Christensen Institute for Disruptive Innovation, Jobs to Be Done Theory, Christensen Institute, 2026, https://www.christenseninstitute.org/theory/jobs-to-be-done/

2. Clayton M. Christensen, Taddy Hall, Karen Dillon and David S. Duncan, Know Your Customers' "Jobs to Be Done", Harvard Business Review, September 2016, https://hbr.org/2016/09/know-your-customers-jobs-to-be-done

3. Tony Ulwick, Jobs-to-be-Done: A Comprehensive Guide, Strategyn, 2025, https://strategyn.com/jobs-to-be-done/

4. Bob Moesta and Chris Spiek, The Four Forces of Progress: Push, Pull, Anxiety and Habit, Jobs to be Done, 2026, https://jobstobedone.org/the-four-forces/

5. Bob Moesta and Chris Spiek, How to Run a Switch Interview, Jobs to be Done, 2026, https://jobstobedone.org/switch-interview/

6. Strategyn, The Innovation Process: A Step-by-Step Guide, Strategyn, 2025, https://strategyn.com/outcome-driven-innovation-process/

7. Carmen Nobel, Clay Christensen's Milkshake Marketing, Working Knowledge, Harvard Business School, 14 February 2011, https://www.library.hbs.edu/working-knowledge/clay-christensens-milkshake-marketing

8. Bob Moesta and Chris Spiek, The JTBD Timeline: First Thought to Purchase, Jobs to be Done, 2026, https://jobstobedone.org/the-timeline/

9. The Re-Wired Group, What is Jobs to be Done? The Complete Guide, The Re-Wired Group, 2026, https://therewiredgroup.com/learn/complete-guide-jobs-to-be-done/

10. Tony Ulwick, Customer Needs Through a Jobs-to-be-Done Lens, Strategyn, 6 November 2019, https://strategyn.com/customer-needs-through-a-jobs-to-be-done-lens/

11. Eric Eskey, Quantify Your Customer's Unmet Needs, Strategyn, 22 October 2019, https://strategyn.com/quantify-your-customers-unmet-needs

12. Strategyn, Find Hidden Innovation Opportunity in the Market, Strategyn, 2025, https://strategyn.com/outcome-driven-innovation/market-opportunity/