Insights / Guide / Consumer behaviour analysis
GuideCustomer interviews that reveal real needs
A dozen well-run interviews often teach more than a large survey. Ask about real situations, listen for workarounds and look for patterns you can then test at scale.
When organisations want to understand their customers, the first instinct is often a survey. Surveys are quick to send, produce charts and feel representative. But they can only answer the questions you thought to ask, in the words you chose. They rarely reveal what you did not know you were missing.
A dozen well-run interviews often teach more. Talking to customers one at a time, about real situations in their lives or work, uncovers needs, frustrations and workarounds no survey would capture. Done well, interviews give you the insight; surveys then tell you how widespread it is. This guide explains how to plan, run and analyse interviews that reveal real needs.
Plan the interviews
Start with a clear question you want to answer. Not “what do customers think of us?” but something specific: “Why do trade customers stop ordering after their first year?” or “How do families decide which energy tariff to choose?” A sharp question keeps the interviews focused and the analysis manageable.
Then decide who to talk to. Aim for variety rather than statistical representation. Include loyal customers, new customers, occasional customers and, crucially, people who left or chose a competitor. If your product has different users and buyers, such as an engineer who uses it and a procurement manager who approves it, speak to both.
- Ten to fifteen interviews usually reveal the main patterns for a single, focused question.
- Interviews of 30 to 45 minutes, by video or in person, work well for most topics.
- A modest thank-you, such as a voucher or donation to charity, improves participation.
- Recruitment through your own customer lists, account managers or a specialist agency each has trade-offs in speed and bias.
Write a short discussion guide: an introduction, five or six topic areas with open questions, and a close. Treat it as a map, not a script. The best insights often come from following an unexpected answer.
A simple structure works for most topics. Open with a few minutes of context about the participant’s role or household, then move to the most recent relevant situation and walk through it in detail. Follow with how they chose between options, what went wrong and what they did about it. Close by asking whether there is anything you should have asked but did not. That last question regularly produces the most useful answer of the session.
Ask about situations, not opinions
The single most important technique is to ask about what people actually did, in specific past situations, rather than what they think or would do. People are poor predictors of their own behaviour and often give answers they believe are expected. Their accounts of real events are far more reliable.
Compare two questions. “Would you use a feature that lets you reorder with one click?” invites a polite yes. “Tell me about the last time you placed a repeat order. What did you do, step by step?” reveals how they really work, where they struggle and what they would value, without you having to suggest it.
Questions that work
- “Tell me about the last time you…” followed by “What happened next?”
- “What were you trying to achieve at that point?”
- “What was the hardest part?”
- “How did you decide between the options?”
- “Who else was involved?”
- “What did you do when that did not work?”
Questions to avoid
Avoid leading questions (“Don’t you find it frustrating when…?”), hypothetical questions about future behaviour (“Would you pay more for…?”) and questions that ask people to design solutions (“What features should we add?”). Customers are experts in their problems, not in your product roadmap.
Let silence work
When a participant finishes an answer, wait a few seconds before speaking. People often fill the silence with the most revealing part of their story. Resist the urge to move on to the next question.
Listen for workarounds
Workarounds are one of the richest sources of insight. When customers describe exporting data into a spreadsheet, keeping a paper list next to the computer, calling a colleague to check something or using a competitor’s product for one particular task, they are showing you a need your product or service does not meet.
A typical example: a distributor interviewing trade customers might find that several keep their own spreadsheet of products they order regularly, copying codes into the online portal each time. No customer asked for saved order lists in a survey. But the workaround makes the need, and the value of meeting it, unmistakable.
Listen too for emotion. Frustration, anxiety, relief and pride mark the moments that matter most to customers. Note what triggered them and ask more about those moments.
Run and record well
Have two people at each interview where possible: one to lead the conversation, one to take notes. With the participant’s consent, record the session so you can return to exact words later. Explain how the recording will be used and stored, in line with your data protection obligations.
Immediately after each interview, spend ten minutes writing down the key points: what surprised you, the main situations described, any workarounds and notable quotes. Memory fades quickly, and these notes become the raw material for analysis.
Watch for your own bias as the interviews progress. After four or five conversations, teams often form a theory and start, without meaning to, steering later participants towards it. Keep asking the same open questions, note evidence that contradicts your emerging view as carefully as evidence that supports it, and invite colleagues who hold different assumptions to observe some sessions.
Treat what participants share with care. Store recordings and notes securely, limit access to the project team, remove names when sharing quotes more widely and delete raw recordings once analysis is complete, unless participants have agreed otherwise.
Look for patterns
Once interviews are complete, bring the team together to analyse them. Write individual observations on cards or digital notes, one per card: a need, a frustration, a workaround, a decision factor. Then group similar cards together and name each group. Patterns that appear across several interviews, especially among different types of customer, are the ones to take seriously.
- Extract observations from notes and recordings, one per card.
- Group similar observations and give each group a descriptive name.
- Note how many participants, and which types, contributed to each group.
- Write each strong pattern as a finding: a need, the evidence for it and why it matters.
- Identify which findings you are confident in and which need further testing.
Be honest about what interviews can and cannot tell you. They reveal what needs exist and why. They do not tell you how many customers share each need. That is the job of the next step.
Test findings at scale
Use the findings to design better quantitative research. A survey written after interviews asks the right questions, in customers’ own language, with answer options grounded in reality. Alternatively, test findings through behaviour: a small change to your website, an offer to a sample of customers or a prototype shown to a wider group.
We had run the same satisfaction survey for years. Twelve interviews told us more about why customers left than all of those surveys combined.
Customer experience director, subscription services client
To begin, choose one specific question that matters to the business now. Recruit ten customers, including at least three who have left or reduced their spending. Write a short discussion guide built around real past situations, run the interviews in pairs over two weeks and analyse them together as a team. You will almost certainly learn something that changes a decision.
Source: Intelliversal analysis.
