Customer interviews that teach you something

Customer interviews are the cheapest research method there is and one of the easiest to get wrong. Done badly, they produce polite agreement with whatever idea you brought into the room. Done well, they tell you how people actually behave, which is far more useful.

What interviews are for

An interview in discovery is not a sales call and not a usability test. Its purpose is to understand the person’s context: what they are trying to achieve, how they do it today, what gets in the way and what they have already tried. You are collecting stories, not votes.

That means the interview is mostly not about your product. If you find yourself explaining what you are planning to build, the interview has turned into a pitch.

Recruiting the right people

The most common reason interviews disappoint is that the wrong people were in them. A few rules of thumb:

• Decide in advance which behaviour qualifies someone. “Has prepared an expense report in the last month” is better than “works in finance”.
• Use a short screener with factual questions, so people cannot simply say what they think you want to hear.
• Include people who stopped using the product, or never started. They know things your happiest customers do not.
• Make it routine. One or two conversations a week, every week, beats an intensive burst twice a year.

Ask about the past, not the future

People are poor predictors of their own behaviour and generous with hypothetical enthusiasm. “Would you use a feature that…?” almost always gets a yes, and the yes means very little. Rob Fitzpatrick’s book The Mom Test is built around this observation and is worth an evening.

The remedy is to ask about specific things that have already happened.

A specific story comes with details that general answers smooth over: the workaround spreadsheet, the colleague who checks everything twice, the deadline that made it urgent. Those details are where opportunities hide.

Running the conversation

Two people from the team is ideal: one leads, one takes notes. Ask for permission before recording. Start broad and let the person talk; follow up with “what happened then?” and “why was that?” rather than moving on to your next question.

Silence is a tool. Most people will fill a pause with something more interesting than their first answer. Resist the urge to rescue them.

Avoid leading questions, and especially avoid describing your solution. If someone asks what you are working on, it is fine to say you will tell them at the end. Usually they have forgotten by then.

Making it useful afterwards

An interview that lives only in someone’s memory is half wasted. Within a day, write a one-page snapshot:

• who the person is, in terms of the behaviour you care about;
• the story they told, in a few lines;
• needs, pain points and workarounds you noticed;
• one or two memorable quotes;
• anything that surprised you.

Snapshots make synthesis possible. After five or six of them, patterns begin to appear: the same workaround mentioned independently, the same moment of frustration. Those patterns feed the opportunity map.

Be careful with the single vivid interview. One articulate customer with a strong opinion can steer a team for months. Look for what repeats, and treat the memorable outlier as a hypothesis to check, not a finding.

Interviews are not the only source

Usage data shows what people do but not why. Support tickets show where people get stuck, at least the ones who bother to write. Sales and support colleagues hear objections every day. Interviews work best alongside these sources, explaining the patterns they reveal.