Interview: preparation and conduct

The interview does not prove demand and does not ask the user to design the solution. His task is to reconstruct the real context: what happened the last time a person used, what stopped them, and what price they already pay for the problem with time, money, or risk.

Why do you need an interview?

Why Interviews Are the Main Discovery Tool

Interviews are conversations with actual or potential users of a product. The goal is to understand their real problems, needs, motivation and context. Unlike surveys, interviews reveal details that people don’t always capture themselves. This is the main way to find real value before the development stage.

When interviews are more useful than other methods

Interviews are in the early stages when you’re not sure what the real problem is or looking for new solutions. For example, a startup is looking for the right audience, an enterprise product is testing new features, and a team is developing an MVP.

Example: Suppose you plan to run a new task for an internal time tracker. Analytics won’t show employees losing motivation because of an uncomfortable interface — but that’s easy to figure out from the conversation.

How to Prepare for an Interview

Define purpose and hypothesis

Before an interview, always formulate one or two key hypotheses: what is important to test? It’s a common mistake to go to an interview for the sake of feeling a product or function, rather than for risk validation.

Example: Hypothesis – employees lack push reminders about deadlines. Find out who should receive them, how often, why current reminders don’t work.

Make a working script.

The script is not a script with exact replicas, but a structure: 4-6 basic questions plus a set of possible clarifications. Questions should be open: not to be imposed by the answer, not to sculpt a function in them.

Anti-pattern: Would you use our service to track time? Tell me how you are tracking your working hours. What is important to you in this process?

Find the right responders

Respondents are as similar as possible to the real or target audience. Representation is important here: don’t just interview friends or colleagues.

Where to look: your databases, profile communities, social networks, sometimes B2B partners. The number of interviews depends on the homogeneity of the segment and the decision to be made. Start with a small series, sort out the discrepancies, and continue until new conversations change a significant part of the picture.

Conducting an interview: details and pitfalls

Right tone and dynamics

Start the interview with warm-up: briefly introduce yourself, explain the purpose, collect permission to record (if necessary). The main thing is to create a friendly atmosphere; do not press, do not rush.

Ask questions consistently, listen more than you say. If a person is confused, calmly return to the topic. Use the paraphrase: Did I get it right that…? – it helps to avoid misunderstanding.

How to capture and process data

Best of all - make an audio recording and simultaneously fix insights with a summary. Don’t try to write down the main thing: problems, patterns, context.

Example: after 8 interviews with accountants, it turned out that they are not tracking work tasks, but projects, so as not to get confused in tax reporting. This led to a rethinking of the product structure.

Anti-patterns:

• You’re using X to confirm questions, right?
• Brainstorm new features instead of deep clarification of problems.
• Interrupting the client with their experience.

Analysis and prioritization of findings

Finding Real Problems

Put aside superficial complaints. At the first stage, look at the frequency – how many people actually face a problem, how important it is, what causes inconvenience.

Frequent criteria

• Some of the respondents themselves came up with workaround solutions.
• The problem pops up without pushing.
• It affects real business metrics (money, time)

Example: if in interviews several executives say that they have chaos because of incorrect deadlines – the problem is clearly not only in the interface.

Let’s take an insight and the next step.

Insight is a short, clearly articulated and verifiable statement of why people behave the way they do. Usually, after the interview, highlight 3-5 main insights that you can discuss with the team.

Next – design hypotheses for the next iteration of Discovery: new interviews, test features, prototypes, surveys.

Resources and useful links

• IDEO Design Kit
• Nielsen Norman Group, Interview Studies
• Product Design Interview Guide by ProductPlan
• Atlassian Guide: Customer Interview for Discovery

FAQ

**Why do you need an interview if you have analysis and feedback? Interviews help to understand the motivations and real practices of users, which are not obvious from the numbers.

**How many interviews do I need to do for Discovery? Usually, 5 to 8 are enough for the first insights, but the figure varies by market and product.

What questions should I ask in an interview? Open questions about current processes, complexities, and life examples. Don’t ask for your opinion on functions.

How can we avoid misrepresentation of data from respondents? Don’t suggest solutions, don’t confirm your ideas, ask for examples and details of their experience.

**What happens if you start giving product advice? Translate the conversation back to his real experience: Tell me the last time you had a situation like this.

**How do you analyze the results of the interview? Highlight common patterns, frequent problems, and credible insights. Compare this with the original hypotheses.