Pebellslearn by building
0%
RoadmapsMatrixCoursesSandboxSign in
Gathering Information  ›  Lesson

Interviews

Interviews 13 minUser interviews are most reliable when they ask neutral questions about specific recent events, rather than asking people to predict their behaviour or agree with an idea.

Interviews

The LunchLoop team is sure it knows why tiffins come back untouched: people forget to skip. That is a guess, and a whole quarter of work is about to rest on it. Six conversations of twenty minutes each can confirm the guess, kill it, or reveal the real cause, but only if the questions are asked the right way.

The rule

Ask about the last real time, not about someday.

People are poor at predicting what they will do and generous with opinions. They are much better at recounting something that actually happened recently. So a good interview is built from past, specific events:

  • Ask about real episodes. "Tell me about the last time your tiffin came back untouched" beats "Would you use a Skip button?", to which almost everyone says yes.
  • Stay neutral. "Isn't it annoying to message support?" plants the answer. "What did you do when your plans changed?" does not.
  • Follow the story. "What happened next?", "Why then?", "Can you show me?" Specific details are where the surprises live.
  • Listen more than you talk. A pause invites people to keep going.
  • Talk to the right people. Subscribers who had untouched boxes, and people who cancelled, not just your happiest customers.
  • Keep numbers small. Five to eight people per group is usually enough. Stop when the stories start repeating.

For a product designer, every interview serves a decision. Here the decision is: "Is forgetting the main cause of waste, or is it something else?"

The shortcut

Run every question through a "last time" filter. Any question that starts with "Would you…", "Do you usually…" or "Do you like…" gets rewritten as "Tell me about the last time…".

Fixing LunchLoop's interview guideStep 1 of 4
  1. "Would you use a one-tap Skip button?"

    Hypothetical, so nearly everyone says yes. Rewrite: "Tell me about the last day you didn't eat your LunchLoop tiffin."

The team's guess was wrong, and the "last time" questions are what exposed it. Hypothetical questions would have confirmed the Skip button and hidden the real problem.

Drills

Answer fast — trust the rule.

quick drill

Which interview question best follows the rule?

quick drill

Why is "Isn't it annoying to message support?" a weak question?

quick drill

A subscriber says, "I'd definitely use a Skip button every week!" How much should this count?

quick drill

After six interviews, four people say their changes come up in the morning, after the cutoff. What next?

Remember

  • Ask about the last real time, not about someday.
  • Shortcut: rewrite "would you / do you usually / do you like" as "tell me about the last time".
  • Trap: leading or hypothetical questions confirm your guess and hide the real problem.

Practice — level up

0/1 solved

Write your answer, then compare it with a rubric and a model answer.

ExplainWhy do people cancel?

Write three interview questions to learn why LunchLoop subscribers cancel their plans, and explain how each one follows the "last real time" rule and avoids leading the person.