How to Prepare for Jobs To Be Done Interviews

A Jobs To Be Done interview reconstructs the story of one real purchase in enough detail that you can see what caused it. Not what the customer likes about the product. What happened in their life that made them say, "today is the day."

Preparing well is most of the work. If you recruit the wrong people, write a rigid questionnaire, or skip the debrief, you get an hour of pleasant conversation and nothing you can act on. This guide covers the four forces behind every purchase, the timeline a buyer actually walks, who to interview, the six-step process, and the questions that get people to tell you the story instead of an opinion.

What is a Jobs To Be Done interview?

The premise of the framework is that people do not buy products, they hire them to make progress. The Clayton Christensen Institute puts it this way: "People don't simply buy or pick products or services; they pull them into their lives to make progress." Jobs To Be Done is the lens that reveals the circumstances driving people toward and away from that decision.

The framework has more than one parent, and it is worth being accurate about that. Tony Ulwick developed Outcome-Driven Innovation in the early 1990s and introduced the underlying idea to Clayton Christensen in 1999. Christensen turned it into a theory of innovation, set out most fully in Competing Against Luck (2016). Bob Moesta, founder and CEO of The Re-Wired Group and a fellow at the Christensen Institute, built the interview method that most practitioners use today, and defines a job as "the progress a person is trying to make in a struggling moment."

For a DTC brand, that makes the interview a causality tool. Analytics tells you what people did on the site. Surveys tell you what people say they prefer. Only the interview tells you which sequence of events produced the order, and that is what you need before you redesign a product page or launch a new line.

The distinction between correlation and causality is the whole point, and it is where most segmentation work goes wrong. As Juliana Jackson put it:

"The single biggest reason most marketing folks mess up customer segmentation: they segment based on correlation. For instance, I'm a female, 30, living in Romania, enjoying hip-hop music and urban fashion. Most people will take those attributes and 'tailor' an experience to sell me an urban fashion backpack. Why is that wrong? There is no causality for me to buy that backpack. Attributes like social demographics, interests, age, etc., can be correlated into associating a customer's interest in a product or a service, but they don't cause that customer to buy."

The four forces of progress

Take a pair of retired parents redecorating their daughter's old room as a guest room. The four forces look like this.

• Push of the situation. The room has been empty since their daughter moved out, and guests keep sleeping on the sofa. The push is the moment the current state becomes a complaint.
• Pull of the new solution. They start seeing bed frames and wardrobes they like, and imagine the finished room. The pull is the appeal of the outcome, not of the product.
• Anxiety of the new solution. Will the bed fit? Is the finish as good as it looks online? What happens if it arrives damaged? Anxiety shows up as a stream of questions about an alternative.
• Habit of the present. As they clear out their daughter's things they get sentimental and start asking whether they need to change anything at all. Habit is the pull of doing nothing.

Most brands spend their budget on pull, because pull is the easy one to advertise. The interview usually shows that the deals are lost on anxiety and habit instead, and those are cheaper to fix: a clearer returns policy, a size guide, a photograph of the product in a real room, a phone number.

The JTBD timeline: how a purchase actually unfolds

The sales pipeline is drawn as a straight line because that is convenient for the seller. The demand side does not work that way. Your job in the interview is to place the customer on this timeline and find out what moved them from each stage to the next.

1. First thought

   The customer thinks about doing something differently. It often starts with an offhand question or a story from a friend. Nothing is being shopped for yet.
2. Passive looking

   They notice things. Their language is still about the problem, not about products. This stage can last months and is invisible in your analytics.
3. The event

   Something happens that turns "someday" into "we should do something about this." A visit is booked, something breaks, a deadline appears. Finding this event is usually the single most valuable thing the interview produces.
4. Active looking

   Now they compare solutions. Language switches from problem to product. Price is usually not the deciding factor yet.
5. Deciding

   Alternatives get weighed and trade-offs get made. Without time pressure, many people stop here indefinitely. This is where anxiety and habit do their damage.
6. Buying and first use

   The order is placed, and the product either delivers the progress or does not. The first week decides whether this customer ever comes back.

Who to interview, and who to leave out

People who considered you and did not buy are a different research project. They can tell you where they stopped, but they cannot tell you what a completed decision looks like, and the story of a decision that never happened has no ending to reconstruct.

RFM segmentation is the practical way to build the list. An RFM analysis scores every customer on recency, frequency, and monetary value, and the customers scoring high on all three, the segment we call Soulmates, are the ones whose jobs you most want to understand. In the Customer Value Optimization model, growth comes from concentrating on the most valuable customers, so the interview list should over-represent them.

Balance that against variety. If all twelve interviews come from the same segment, the same age group, and the same city, the patterns you find will be real but narrow. Nexus by Omniconvert builds the RFM segments from your order data, so you can pick a spread of profiles from within the segments that matter rather than guessing from a spreadsheet.

The six-step JTBD process

1. Frame the question

   Decide what you are trying to find out: what causes people to say "today is the day I need this." Write down every question you would like to ask, then find the one or two that would answer most of the others. Resist the researcher's trap of stuffing thirty questions into an hour; a long list produces short answers.
2. Set up the interviews

   Recruit customers who made progress, meaning people who bought. Aim for a wide spread of ages, genders, locations, and order sizes, weighted toward your highest-value RFM segments. Interview recent buyers where you can, and offer a real incentive; you are asking for an hour of someone's attention.
3. Conduct the interviews

   Plan for 45 to 60 minutes with two people on your side: one leading, one listening. Keep the customer talking about the progress they were trying to make rather than about your product's features. Treat the script as a map, not a questionnaire, and follow the story where it goes.
4. Find the patterns

   Debrief within the hour, for 30 to 45 minutes, while the story is still fresh. Argue about what you heard, and settle on the customer's own words rather than your paraphrase. Around 10 to 12 interviews are usually enough for patterns to repeat.
5. Detail the clusters

   Group interviews that tell similar stories. Each cluster is a pathway by which people make progress, and each one gets described in the same terms: context, struggling moment, pushes and pulls, anxieties and habits, desired outcome, hire and fire criteria, trade-offs, and the baseline quality expected.
6. Prototype opportunities

   Turn each job into something you can ship and measure: a change to product page copy, a bundle, a guarantee that kills a specific anxiety, a new product. Most brands only ever look at the supply side. The clusters give you the demand side, which is where the unclaimed opportunities are.

Bob Moesta's interview technique borrows from investigative interviewing, and several of the moves feel odd until you try them. He teaches them in the CVO Course:

• Keep the script as a framework and let the dialogue run naturally.
• Listen actively for changes in tone; hesitation marks the interesting parts.
• Play confused on purpose to make the customer supply the full context.
• Play the story back in the wrong order, or slightly wrong, so the customer corrects you and adds detail.

Set expectations at the start. Tell the customer you want to understand the language people use around buying this kind of product, that it is closer to a documentary than a survey, and that some questions will sound strange.

Questions to ask in a JTBD interview

These are our own starting questions, not a fixed script. Use whichever ones the story calls for, and skip the rest.

• First thought. When did you first think about solving this? Where were you? What made you think about it that day? Had you thought about it before and let it go?
• Passive looking. What did you do about it back then, if anything? What did you notice that you had not noticed before? Who did you talk to about it?
• The event. What changed that made this urgent? Was there a date you were working toward? What would have happened if you had done nothing?
• Active looking. What else did you look at? What did you type into the search box? What did you almost buy, and what stopped you? Who else had a say?
• Anxieties. What worried you about ordering? What did you check before you bought? What would have made you abandon the order?
• Habits. What were you using before? What did you have to give up? What made it hard to let go of the old way?
• The purchase. Walk me through the day you ordered. What time was it, where were you, what device? How long between deciding and paying?
• First use. What happened when it arrived? What was the first week like? What did you tell other people about it? What has changed for you since?

Two habits carry most of the quality. Ask for the specific occasion rather than the general habit ("tell me about the last time" beats "what usually happens"). And when an answer sounds like a slogan, ask what happened just before, until you get back to an event.

How to debrief and cluster the interviews

Give every debrief file the same six sections so that clustering later is mechanical rather than interpretive.

• Setup: the context, who the customer is, what their life looked like before.
• Pushes: the problem or pain point that started the process.
• Pulls: what attracted them to the new solution.
• Habits: the resistance to change and what they had to give up.
• Anxieties: the questions and worries about the product.
• Observations: anything true and interesting that does not fit the other five.

Once the clusters are stable, each one should be readable as a job: a situation, a struggle, and the progress the customer wanted. The table below shows how to read each signal and what to do with it.

Jobs To Be Done interviews are the third step of the Customer Value Optimization timeline, and they only pay off if the output leaves the research document. Every cluster should end as a hypothesis with an owner and a metric.

Frequently Asked Questions

Alexandra Panaitescu

Content Marketing Specialist

Alexandra Panaitescu is a B2B content marketing specialist with over 8 years of experience building data-driven content strategies and inbound campaigns that help businesses grow, from generating qualified leads to establishing brand authority and revenue.