left arrow icon

Run jobs-to-be-done interviews with recent adopters

A jobs-to-be-done interview rebuilds how a customer came to switch: the trigger, what they used before, and the progress they wanted. Practice says to interview recent adopters, since habitual users have no energy left in the decision. thoughtbot typically needs 8 to 10 interviews. At Commoncog, interviews revealed two buying paths, and data showed 56.5% of members came by the one the founder had not assumed.

Evidence C, limited evidenceStrong when it fits

Stage
Stage 2. Users, no retention signalDo the first users get value, and do they come back?
Category
Customer research
Time
1 to 2 weeksAwiser's estimate of the active work
Cash
No cash
Live calls
Needs live calls
First result
Within weeks
Risk
LowThe cost is recruiting and interview time. The risk is interviewing settled users and writing down a job nobody hired you for.
Needs
Customers who adopted within about the last six months and stayedA live conversation with eachSignup or purchase data to check what you hear

When does it fit?

  • You have customers who adopted recently: Bob Moesta wanted people who had recently onboarded rather than people using the product every day.[1]
  • You want to know what people switched from: Moesta says they always switch from something.[1]
  • Your growth model rests on assumptions: Commoncog's interviews showed its causal model of growth was wrong.[4]

When should you skip it?

  • Your only interviewees are long-time daily users: Moesta says habitual users make poor JTBD interview subjects.[1]
  • You can only ask in writing: Cedric Chin says customers forget how they bought and invent reasons that the interviewer has to catch.[4]

How do you run it?

  1. Pick recent adopters

    Customers who started within about six months and stayed. Add people who switched away, if you can.

  2. Rebuild the timeline

    When they first thought about it, what they were using, what happened that day, and what else they considered.

  3. Probe the reasons

    People invent reasons after the fact. Ask what happened just before, and for specifics.

  4. Name the progress they wanted

    What "better" meant to them, in their words.

  5. Check against your data

    Test the paths you hear against signup or purchase data.

  6. Allow more than one job

    If timelines differ, you may serve two jobs or two segments.

What usually happens?

  • A small number can be enough: thoughtbot typically needs about 8 to 10 switch interviews, and Moesta compares ten interviews to a thousand surveys.[1][2]
  • It is hard: Cedric Chin of Commoncog says the first thing you discover about the JTBD interview is how hard it is to do.[4]
  • You may find more than one job: Commoncog found that members who came through different paths experienced different timelines.[4]

Who has done it, and what happened?

Each case is one team's own report: what happened to them, not a typical result.

  • Cedric Chin, founder of Commoncog, a paid membership publication2025, Positive result

    Ran jobs-to-be-done interviews with paying members about how they came to buy, then checked the paths against membership data.

    Cedric Chin reports that 56.5% of members came through the second path and only 43.5% were newsletter subscribers before purchasing.[4]

    A publication rather than a software product, the number of interviews is not stated, and the outcome is a corrected growth model, not a measured change in revenue.

What does the evidence say against it?

  • Old customers make poor subjects: Commoncog ended up with useless transcripts from customers who bought too long ago.[4]
  • No dataset links jobs-to-be-done interviews to a measured change in conversion or retention, and one recent case was found.

How do you read the result?

What to count
Interviews with recent adopters, and the triggers, earlier solutions and wanted progress that repeat
When to read it
After eight to ten interviews
How to decide
Rewrite positioning and onboarding around the job that repeats, and check it against your data. Cedric Chin advises starting with customers who purchased less than six months ago.[4]

What goes wrong?

  • Forcing one job: the Clayton Christensen Institute notes that people hire products for different jobs as their circumstances change.[3]
  • Taking stated reasons at face value: Cedric Chin warns that people make up reasons and you have to catch them.[4]
  • Expecting everyone to accept: Clozd, a win-loss vendor, sees 30 to 40% of customers who bought accept an interview when they know who is asking.[5]

How was it graded?

Evidence
C. Limited evidenceJobs-to-be-done interviewing is a recognised method, and one recent documented case shows it correcting a founder's growth model, but no dataset measures its outcomes.
Verdict
Strong when it fitsWorth it once you have customers who adopted recently: it shows why they switched, which shapes positioning and onboarding. With only long-time users, it teaches little.

How the grades and verdicts are decided

Sources

  1. 1
    Bob Moesta on Jobs-to-be-Done (opens in a new tab)Intercom, Des Traynor, with Bob Moesta. May 12, 2016. Method.
    In the case of people who use Intercom, you wanted to talk to people that'd recently onboarded or recently signed up rather than people who used the product every single day, because there's no energy involved in their hiring.
    They always switch from something.
    This is why habitual users of a product don't make great JTBD interview subjects. They aren't looking to improve their situation.
    We'll do ten interviews, but it's like having a thousand surveys.
    Read on October 20, 2026.
  2. 2
    Switch Interviews (opens in a new tab)thoughtbot. Undated. Method, from a seller of what it measures.
    Typically it takes about 8-10 interviews to get valuable information, but we have gotten away with less.
    Read on October 20, 2026.
  3. 3
    Jobs to Be Done (opens in a new tab)Clayton Christensen Institute. Undated. Method.
    We hire for different Jobs based on our changing circumstances.
    People don't simply buy or pick products or services; they pull them into their lives to make progress.
    Read on October 20, 2026.
  4. 4
    Putting the Jobs to be Done Interview to Practice (opens in a new tab)Commoncog, Cedric Chin. July 9, 2025. Founder report.
    That meant that our causal model of Commoncog's growth was wrong.
    First, most customers can't remember how they bought your thing. Second, the human mind is not designed for self knowledge.
    Humans will make up reasons for why they do things, and you have to catch them when they do.
    The first thing you'll discover about the JTBD interview is that it's really hard to do.
    During the JTBD interviews, we began to realise that members who became paying customers through the first path experienced a different timeline from readers who became paying customers through the second path.
    The most basic one is to never interview a customer who bought too long ago. At Commoncog, we didn't take this seriously enough, and so we ended up with a bunch of interview transcripts that were useless for our purposes.
    This leads to a counter-intuitive recommendation: at the beginning of your JTBD journey, limit your interviews to customers who have purchased less than six months ago.
    We learnt that 56.5% of members came from the second path. Only 43.5% were newsletter subscribers before purchasing.
    Read on October 20, 2026.
  5. 5
    Selecting Buyers for Post-Decision Interviews (opens in a new tab)Clozd, Jonathan Stevens. April 20, 2026. Vendor data, from a seller of what it measures.
    Non-Blind Wins: 30% - 40% participation
    Read on October 20, 2026.

Sources last checked on October 20, 2026. First published on October 20, 2026.

Awiser is the builder network where founders show their skills, find collaborators, and grow their projects with an AI copilot.