The interview is the easiest research tool to organize and the easiest to get wrong. One badly phrased question is enough for the other person, out of politeness or laziness, to say what they think we want to hear. The result is a set of kind opinions that confirm our ideas and teach us nothing.
What an interview is for
An interview is not for asking people what they want. It is for understanding how they experience a problem: what they do today, what tools they use, where they lose time, what annoys them, what workarounds they have invented. People are excellent witnesses of their own experience and poor designers of their ideal product. A phrase often repeated among researchers sums it up: watch what people do, not only what they say.
Past events, not future hypotheticals
The most important rule is to ask about concrete episodes that have already happened, not hypothetical behavior.
| Weak question | Better question |
|---|---|
| «Would you use an app to split expenses with your flatmates?» | «Tell me about the last time you split a bill. How did you do it?» |
| «Would you like to get a reminder?» | «Have you ever forgotten a deadline? What happened next?» |
| «How often do you shop for groceries online?» | «Think of your last grocery shop. Where, how and why there?» |
Almost everyone says yes to a hypothetical question, because it costs nothing. The story of a real episode, instead, reveals details nobody would have imagined: the shared spreadsheet nobody updates, the group chat used as an archive, the phone call to the bank to check a payment.
Open and neutral questions
A closed question gets a yes or a no. A leading question already contains the answer. Compare:
- «Do you find the booking process difficult?» suggests that it is.
- «How did it go the last time you booked?» leaves room for any answer.
Other useful prompts to keep the story going without steering it:
- «What happened next?»
- «Can you give me an example?»
- «Why does that matter to you?»
- «How did you realize there was a problem?»
Prepare a guide, not a script
An interview guide is a list of themes and starting questions, not a questionnaire to read line by line. It usually includes:
- a short introduction to the purpose and duration;
- questions about context, for example work or household if relevant;
- the story of one or two recent episodes related to the topic;
- follow-ups on the critical points that came up;
- a close in which you ask whether there is anything you haven't covered.
A forty-minute interview with ten well-chosen questions yields more than one with thirty rushed ones. The best moments almost always come from an unexpected answer worth following.
The value of silence
Interviewers tend to fill every pause. That is a mistake: after a short answer, waiting a few seconds in silence often prompts the other person to add the most interesting detail. Silence conveys attention, not awkwardness, when it comes with a relaxed attitude.
Likewise, avoid reacting with enthusiasm or disapproval. An over-emphatic «Oh, interesting!» steers the following answers toward that topic.
Common mistakes
- Presenting your idea too early. Once the prototype is shown, the conversation shifts to judging the prototype and the story of the problem stops.
- Asking two questions in one. «How do you book and how much does it cost you?» gets an answer to only one of them.
- Interviewing only people who agree. Satisfied customers are valuable, but those who left the service often teach you more.
- Asking what price they would pay. Hypothetical answers about money are among the least reliable.
Taking notes and analyzing
During an interview it is hard to listen and write at the same time. If possible, work in pairs: one leads and the other takes notes, separating verbatim quotes from interpretations. With the person's consent you can record the audio.
After five or six interviews, gather the observations on sticky notes or in a shared document, one per line, and group them by affinity. Recurring themes emerge: needs, obstacles, habits. Each theme should be backed by several concrete episodes, not by a single striking sentence.
In short
- An interview is for understanding how people experience a problem, not for asking what they want.
- The best questions are about real, recent episodes, not hypothetical behavior.
- Open, neutral questions and well-handled silences bring out the most useful details.
- Show the prototype at the end, or in a separate session, so as not to steer the story.
- Analysis groups observations into themes backed by several concrete episodes.
