Jobs to Be Done Interview: Screener, Sample and Script

A jobs to be done (JTBD) interview, also called a switch interview, is a one-to-one conversation that reconstructs, in time order, how someone came to hire a product or service to make progress in their life. The core recruit is a person who switched to or from a product 30 to 90 days ago. You do not interview your happiest customers. The originators of the method say to recruit switchers, and to avoid "best customers" as a category.

In the interview you map four forces:

• Push: what made the old way unacceptable.
• Pull: what attracted them to the new option.
• Anxiety: what made them hesitate.
• Habit: what kept them with the old way.

A second, optional segment is people who felt the struggle but did not switch: they seriously considered a change and kept the old way. The originators describe this as non-consumption, where solutions seem too expensive, too complicated or out of reach, and note that it is often the largest untapped market (The Struggling Moment). Treat it as its own segment with its own screener wording, below. Do not mix it into the switcher group.

Plan for 45 to 90 minutes per session. Patterns for a given job often start to appear within 5 to 8 interviews, according to the people who developed the switch interview. Our recommendation is 8 usable interviews per segment, scheduling 11.

Our reasoning:

• Eight is the top of the range at which patterns often appear, so it leaves room for one or two stories that turn out to be unusual.
• Keep going to around 12 if the last two interviews still turned up a new push, a new anxiety or a competitor you had not heard of. Stop when they do not. That stopping point is saturation, the point where new interviews stop turning up new forces. Sample size in qualitative research explains why saturation, not a fixed number, is the test.
• Run each segment separately. If first-time buyers and repeat buyers have different struggles, eight blended interviews will describe neither group well.

Keep sessions to 90 minutes at most. People take time to relax before the real timeline surfaces, so 45 minutes is the minimum.

Use a six-question screener, a short set of questions that decides who qualifies, that tests for a dated, personal switch 30 to 90 days ago and a story the person can tell. Book only if questions 1 to 4 and 6 pass. Question 5 is a flag: if the answer is flat, ask one follow-up about the day they started looking, then pass or fail. For a general-purpose version, see our research screener template.

The window follows the originators' guidance: switchers from the last 30 to 90 days, because beyond about 90 days the details you need are gone. Our reading of the lower bound is that someone who switched last week has barely lived with the choice. Treat under 30 days as a fail.

Question 5 is a quick test for a real struggle. A flat answer means the struggle was small or you have not found it yet, according to the originators, so push back in time before you decide.

Question 3 matters most where a purchase is decided jointly. Where a household shares a decision, we would book the person who paid and ask who else weighed in. If you can add an interview with that person, do. Do not substitute them.

Optional segment: people who kept the old way. Replace questions 1 and 2:

• 1b. "In the last 30 to 90 days, did you seriously consider changing how you handle [the job] and decide to keep things as they were?" Pass: yes, and they name what held them back.
• 2b. "When did you last decide to keep things as they were?" Pass: a date 30 to 90 days ago.

Questions 3 to 6 stay the same.

To book 11 sessions you will need to screen about 22 people if half of them qualify, and about 110 if one in ten does. The tighter the behaviour you need, the more people you have to screen. Work it out before you promise a date.

This is our model. The inputs are illustrative assumptions, not measured rates. Replace them with your own after a pilot.

Step 1: sessions to schedule. Usable interviews wanted ÷ (show-up rate × usable rate).

• Inputs: 8 usable interviews, 85% show up, 90% of those are usable (the recording works, the person is the decision-maker, the story is complete).
• Working: 8 ÷ (0.85 × 0.90) = 8 ÷ 0.765 = 10.5, rounded up to 11 sessions.

Step 2: people to screen. Sessions to schedule ÷ pass rate.

Step 3: people to contact. People to screen ÷ the share of those contacted who answer the screener. We assume 25% here.

The working: 11 ÷ 0.50 = 22; 11 ÷ 0.30 = 36.7, rounded up to 37; 11 ÷ 0.10 = 110. Then 22 ÷ 0.25 = 88; 37 ÷ 0.25 = 148; 110 ÷ 0.25 = 440.

If you have a customer database with an order date field, you can email recent buyers directly. If you do not, which is usual for a new category, a competitor's customers or a segment you have never sold to, you are recruiting from outside your list, and the pass rate is the figure to ask your recruiter about first.

Open on the day they bought, to anchor the story in concrete facts, then move forward in time from the first thought to the new reality, using the five blocks below. Always ask about specific events, such as the last time they did it, rather than how they usually do it. Nielsen Norman Group makes the same point about all interviews: jog memory with specific events, not general processes.

This pacing is our recommended run sheet. The block names follow the timeline of the switch interview. The minutes are ours; for 60 to 90 minutes, stretch blocks 2 to 4.

Total: 8 + 10 + 12 + 10 + 5 = 45 minutes.

Use pauses and do not rush to fill silence. Reflect back what you heard and say "tell me more about that" when something interesting appears. Interview in pairs where you can: one person leads the conversation and the other listens for gaps in the timeline and for the four forces.

A worked case shows why the timeline matters. In a recorded switch interview, a mattress purchase looked like an impulse buy on a shopping trip. Reconstructing the Saturday, the year of poor sleep, the three months of online research and the worry about mattress shops showed a long struggle behind a short purchase. The interview ran about 45 minutes.

Jobs to be done interviews fail mostly at recruiting and at asking "why". We recommend treating these five as the main risks, each with a fix you can apply before the first call.

• Wrong person. In a purchase involving several people, an interview with someone who did not decide produces a story about someone else's decision. Fix: question 3 of the screener.
• Too long ago. Beyond about 90 days the details you need are gone, according to the originators. Fix: question 2, and an order record.
• Asking why. A direct "why did you switch?" tends to bring a tidy reason, such as price or features. A dated timeline of events brings the situation behind it. Fix: ask what happened first, next and then.
• Reported, not observed. Interviews capture what people remember doing, not what they did, as Nielsen Norman Group notes. Fix: ask them to check an order email or app history, and compare against your own sales data.
• Doing it alone. The method is demanding. Fix: pair up, or bring in a moderator. Our in-depth interview method guide covers briefing one.

You send us a brief with your screener, segment and city, and we tell you whether the switchers you need can be found. We then recruit to your screener and run or support the sessions.

How it works:

1. You share the research brief: your objective, the behaviour that qualifies someone (the screener above), the segment, geography, sample size and timeline.
2. We assess feasibility and send a tailored proposal and quote. We do not work from a fixed rate card because the price depends on method, audience, recruitment complexity and scope.
3. We recruit by city, check participant identity before invitation, and can run sessions in languages other than English. We recommend interviewing in the language the participant used when making the decision.
4. We schedule the sessions. A moderator can be supplied, or your team can lead; we agree this in the scope.
5. You receive the deliverables we agree before the study begins, which can include recordings, transcripts, structured data exports and a report.

We also offer a free pilot, so you can test your screener and measure your own pass rate before committing to a full study.

Consent matters because these sessions are usually recorded. Tell participants before they start what is recorded, who will see it, and how they can withdraw. India's Digital Personal Data Protection Act, 2023 is the law on digital personal data. As of 14 October 2026, the Digital Personal Data Protection Rules, 2025, notified on 13 November 2025, are coming into force in phases, with the main obligations applying from 13 May 2027 (Press Information Bureau). We are engaging a specialist to assess and establish our compliance framework under the Act. Our informed consent guide gives a form and the Indian rules. We are guided by the Market Research Society of India and by ESOMAR, the global body for market, opinion and social research, and its code of conduct, on consent and confidentiality.

We aim to reply within one business day (Monday to Friday, 10:00 to 18:00 India Standard Time); detailed briefs may take longer. To start a free pilot of your screener, email your brief to support@qualfacto.io or message our team on WhatsApp.