Customer Interview Guide

A jobs-to-be-done style interview guide for one research question: an opening that stops people performing, questions that get stories instead of opinions, the probes to use when an answer goes abstract, and the questions to never ask.

A Markdown interview guide containing: (1) a framing script to read verbatim; (2) four question blocks — Context, The last time, The decision, The alternatives — with three to five past-tense open questions each and a minute budget; (3) probing follow-ups for abstract answers; (4) a closing question; (5) a 'Do not ask' list of four or five questions each with the reason it produces unusable data.

# Interview guide — abandoned onboarding (30 minutes)

## Framing script (read this, 2 minutes)
> Thanks for making time. To be upfront: I am not going to try to get you back, and nothing you say will end up in a sales email. We know you signed up and did not finish setting things up, and I want to understand what was going on for you at the time — including if the honest answer is that something else came up and you forgot about us. That is a genuinely useful answer and I would rather have it than a polite one. Mind if I record so I am not typing?

## Block 1 — Context (6 minutes)
- Tell me about your role and what your team is responsible for.
- What was going on that led you to look for something like this?
- Who else was involved in that, if anyone?
- What were you using before, or instead?

## Block 2 — The last time (8 minutes)
- Take me back to the day you signed up. What were you doing just before?
- Walk me through what you did after you created the account, as far as you can remember.
- Where did you get to before you stopped?
- What was happening around you at that point — what else was on that day?
- When did you next think about it, if you did?

## Block 3 — The decision (8 minutes)
- What made you stop when you did?
- Was that a decision, or did it just not continue?
- Did you look for help at that point? What did you do?
- What would have had to be different for you to keep going that day?
- Did you tell anyone on your team about it?

## Block 4 — The alternatives (5 minutes)
- What did you end up doing about the original problem?
- Did you try anything else after us?
- How is that working out?
- If the problem is still there, what are you doing about it today?

## Probes for when an answer goes abstract
- "Can you give me an example of that?"
- "When did that last happen?"
- "What did you actually do?"
- "Who else was in the room when that happened?"
- Silence. Count to five before filling it. The second half of an answer is usually the true half.
- "You said it was fiddly — what specifically was fiddly?" Repeat their exact word back rather than paraphrasing it.

## Closing question
- Is there anything I should have asked you and did not?

## Do not ask
- **"Would you use it if we fixed the data import?"** — asks for a prediction about a hypothetical. People say yes to be kind and the answer has no information in it. It also tells them your hypothesis, which contaminates everything after it.
- **"How important is onboarding speed to you, on a scale of one to five?"** — everyone says four or five. Importance ratings in an interview measure agreeableness, not priority.
- **"What features would make you come back?"** — invites them to design your product, and people are reliably bad at it. Ask what they did, not what they want.
- **"Did you find the data import difficult?"** — leading and answerable with a yes. If step 4 is really the problem it will surface unprompted in block 2, and if it does not surface unprompted, that itself is your finding.
- **"Were you happy with your experience?"** — invites a verdict rather than a story, and the verdict will be softened because you are on the call.

_One note on your hypothesis: this guide deliberately never mentions the data import. If step 4 is the cause, several of twelve interviewees will raise it in block 2 without prompting. If nobody does, you have learned something more valuable than a confirmation._