> ## Content Index > Fetch the complete content index at: https://pulsecorp.co/blog/llms.txt > Use this file to discover other available public pages before exploring further. # How to run jobs to be done interviews at scale - URL: https://pulsecorp.co/blog/jobs-to-be-done-interviews/ - Published: 2026-11-03T12:00:00.000Z - Updated: 2026-11-03T12:00:00.000Z - Description: Jobs to be done interviews at scale: how to keep hundreds of interviews on track, sort what buyers say and check each reason across groups. - Author: Khyati Ojha A manager at the car company where you work asks you a question. Why did people buy its first electric car? She wants the reasons from hundreds of recent buyers. So can a careful interview work with that many people? The interview she means is a jobs to be done interview. Jobs to be done is the idea that people buy a product to get a job done, like buying an alarm clock to wake up on time. The interview asks one buyer to tell the story of one purchase. Jobs to be done interviews at scale work if every interview follows the same story and asks the same follow-up questions. You also read the first few before you run the rest. Then you check each reason across groups of buyers before you trust it. Each part of the interview needs one small change when you go from a few talks to hundreds. ## The short version - Turn each stop of the buyer's story into a question every interview must reach. - Ask about dates and events, because people remember decisions through them. - Talk to buyers soon after they buy, while the details are still fresh. - Write down the vague answers that always get a second question. - Read the first few interviews and fix the questions before you run the rest. - Keep a quote on every reason, and check it across groups of buyers before you trust it. ## What does the interview listen for? A jobs to be done interview listens for the story of one decision, told in order. It starts with the first thought that the old way wasn't working. It ends with the first weeks of using the new thing. Researchers call this story the timeline. For a car buyer, it usually has five stops: - The first thought, when the old car starts to feel like a problem. - Passive looking, when they notice electric cars on the road but don't search yet. - Active looking, when they read reviews, compare prices and book a test drive. - Deciding, when they pick a car and pay for it. - First use, when they live with the new car for a few weeks. The "job" is the progress the buyer wanted to make. For one driver, it might be a cheaper, calmer drive to work. Along the way, four forces move the buyer. Two of them move the buyer toward the new car, and two hold them back. Here is how each one might sound at the car company: - Push is what makes the old way feel bad. The fuel bill kept going up. - Pull is what makes the new thing look good. Charging at home overnight sounds easier than a trip to the fuel station. - Anxiety is a worry about the new thing. What if the battery runs out on a long trip? - Habit is the comfort of the old way. They always filled up at the same station on the way to work. Buyers like to tell a short, neat story about why they bought. So a good interviewer keeps bringing them back to real dates and events, and that takes patience. With hundreds of interviews, one person can't do that every time. So the interview guide, the list of questions every interview follows, has to do that job instead. ## How to keep hundreds of interviews on track With a few interviews, one skilled person keeps each talk on track. With hundreds, nobody can sit in on all of them. Each part of the interview slips in its own way, and each part has a simple fix. | Part of the interview | With one interviewer | With hundreds of interviews | What you change | | --------------------- | ------------------------------------------------------- | ------------------------------------------------ | ----------------------------------------------------------- | | Recruiting | A few recent buyers are easy to find | Buyers from months ago tell a neater story | Invite buyers in waves, soon after they buy | | The timeline | The interviewer brings people back to the first thought | People jump to the reason they tell friends | Story steps every interview must reach | | The second question | The interviewer hears a vague answer and asks more | Follow-ups change from one interview to the next | A list of vague answers that always get a follow-up | | Analysis | One person remembers every interview | Nobody can remember hundreds | A shared list of labels built from the first few interviews | | Reporting | A few clear stories | Stories get blended into one dull line | Every label keeps its quotes | ### 1\. Recruit recent buyers in waves Recruiting means finding and inviting the people you will interview. A buyer from last month remembers more than a buyer from last year, because details fade. With hundreds of interviews, you run out of recent buyers fast. So invite people in small groups, called waves, as they buy. Each wave then talks to you soon after its purchase. ### 2\. Turn the timeline into the guide Put each stop of the timeline in the guide as a question every interview must reach. If you skip this, many interviews start at the purchase and miss the first thought. Tie each question to a moment in time. You might ask "When did you first look at electric cars, and what was going on in your life then?" People remember decisions through events. So a question tied to a date or an event, called a dated anchor, brings back details that "Why did you buy it?" can't. ### 3\. List the vague answers that always get a follow-up Some answers sound complete but explain nothing, like "it was time" or "it just made sense". Write them into the guide, each with its own follow-up question. A follow-up like this is called a probe. It's a short question that asks someone to say more. Say a buyer tells you the electric car "just made sense". You ask when it started to make sense. She says her new job came with a much longer drive, and her fuel bill went up fast. That's a push you would never hear without the second question. A good probe never hints at the answer you hope for. There's a guide to [judging a good probe against a leading question](https://pulsecorp.co/blog/good-probe-or-leading-question-how-to-judge-ai-interview-quality-in-one-transcri/) if you want to check yours. ### 4\. Read the first batch before you launch the rest Run a small first batch of interviews. Then read every transcript, the written record of each talk. Check that each interview reached the first thought, that the follow-ups stayed fair and that people understood the dated questions. Fix the guide, and then launch the rest. If you skip this step, one mistake in the guide repeats in every interview. ### 5\. Spend your own time where it counts At hundreds of interviews, you stop running each one yourself. So your time goes to the parts that need your judgment. You choose which decision to study and write the dated anchors. Then you read the first batch and decide what the forces mean. Some reasons for buying are private, like a new baby or money worries. So decide before you start which groups you will look at most closely. Then listen to some recordings from each of those groups. AI-moderated means an AI asks the questions, while human-moderated means a person does. It helps to know [how AI-moderated and human-moderated interviews differ](https://pulsecorp.co/blog/ai-moderated-interviews-vs-human-moderated-interviews-whats-the-difference/) before you plan where your own time goes. ![Table graphic of five interview parts, from recruiting to reporting, showing what breaks at hundreds of interviews and the change that fixes each.](https://pulsecorp.co/blog/content/images/2026/11/jobs-to-be-done-interviews-q35306.webp) ## How to sort hundreds of answers Once the interviews come in, you need a way to sort what people said. Tagging each answer with a label is called coding, and each label is a code. Start with a short list of codes from the first batch, one for each push, pull, anxiety and habit you heard. That list is a forces codebook. A [codebook](https://pulsecorp.co/glossary/codebook/) gives each code a plain meaning, so everyone tags the same kind of answer the same way. Use the buyers' own words, and add a new code only when a new force turns up. Keep at least one quote, with a link to its transcript, on every code. Then anyone can check that a code like "battery worry" means what it says. It's the same way you would [audit each theme back to the exact quote](https://pulsecorp.co/blog/synthesis-you-can-trust-how-to-audit-an-ais-interview-themes-back-to-the-exact-q/). Code the trigger separately from the forces. The trigger is the event that started the story, like a new job, a rise in fuel prices or a neighbor's new car. It tells the marketing team when people start looking, so it's one of the most useful codes you have. Try not to squash every reason into one job. Sometimes two groups of buyers switched for different reasons. Then keep two jobs, one for each group. Next, check each force across segments before you call it a finding. A segment is a group of buyers you look at on its own. At the car company, useful segments might be: - Each country where the car sells. - First-time electric buyers, and people who owned one before. - Drivers who get the car from work, and people who buy their own. Keep interviewing in a segment until new interviews there stop adding new forces. Say home charging comes up as a strong pull. When you split the answers, it pulls buyers who have a driveway. But it worries buyers who live in apartments, because they have nowhere to plug in. Both are true, and they lead to different plans. Interviews explain why people switched, but they can't count how many feel the same way. To size each force, follow up with [a survey to count how many share each force](https://pulsecorp.co/blog/surveys-tell-you-what-voice-interviews-tell-you-why-the-question-your-nps-score/). Sizing a force means estimating how many buyers share it. If a force shows up in every segment, it belongs in the main plan. If it shows up in only one, plan for that segment and say so. ![Diagram of home charging pulling buyers with a driveway but worrying buyers in apartments, with the rule: a force in every segment goes in the main plan, a force in one segment gets its own plan.](https://pulsecorp.co/blog/content/images/2026/11/jobs-to-be-done-interviews-y0ydut.webp) ## Where Pulse fits Jobs to be done interviews at scale need one guide that every interview follows, and quotes that stay with every code. [Pulse runs AI-moderated interviews built from a brief](https://pulsecorp.co/product/ai-interviews/#:~:text=AI%2Dmoderated%20interviews%20run%20side,people%20you%20can%20reach.), the short note that says what you want to learn and who to ask. Jobs to be done is one of the study types it runs. The AI moderator asks every buyer the same questions from one guide. Its follow-ups build on what the buyer said earlier, and a vague answer gets a second question. So "it just made sense" gets explored in every interview. One study can run across [several markets from one guide](https://pulsecorp.co/product/multi-market/), in [240+ languages, on request](https://pulsecorp.co/product/languages/#:~:text=Pulse%20runs%20AI%2Dmoderated%20interviews,several%20languages%20at%20once.). Pulse groups answers into themes, each with the quotes behind it, and every quote traces to its transcript. Results are ready within minutes of the last interview. You keep the parts that need your judgment: the decision to study, the dated anchors and what the forces mean. To see how a study like this would run, [book a demo](https://pulsecorp.co/demo/). ## Questions readers ask ### 1\. What does a jobs to be done report look like with hundreds of interviews? It's a short list of jobs. Each job shows the trigger that starts it, the forces behind it and how those forces differ between segments. Every force keeps two or three quotes in the buyers' own words. Once a follow-up survey has run, each job also shows how many buyers share it. ### 2\. How do jobs to be done interviews fit into UX research? UX research studies how people choose and use a product. Jobs to be done interviews explain why people choose it in the first place. Usability tests show how they use it once they have it. So the interviews come first, when the team is deciding what to build or who to sell to. Usability tests come later, once there's a design to try. ### 3\. Should you interview people who looked but didn't buy? Yes, a small group of them is worth adding. People who test-drove the electric car and then bought something else show you which worries and habits won. Their stories often point to the fix that turns a near-buyer into a buyer, like a clearer answer about charging on long trips.