How to Conduct Your First Customer Interview

The first customer interview is where most student founders discover that their idea and reality are two different objects. It is also the step most often skipped, because it feels awkward, unscientific and slow compared with building something.

It is none of those things. An interview is a structured conversation with one goal: to learn what a person actually did, in the past, when the problem you care about occurred. Everything else — your idea, your solution, your enthusiasm — is a contaminant.

Find the right ten people

You do not need a hundred interviews. Ten well-chosen conversations will change your thinking more than a survey with three hundred responses.

The selection criterion is simple: the person must have experienced the problem recently. Not "might have". Not "would probably relate to it". Recently, in a way they can remember and describe.

Places students can realistically recruit from:

• Their own department, hostel, club or team.
• Alumni one to three years ahead, who remember the problem and are now reachable as adults.
• Local businesses within walking distance.
• Online communities organised around the activity, not around startups.
• Second-degree contacts: ask each interviewee for one introduction at the end.

Keep the ask small and honest. "I'm a student researching how fest teams manage sponsor conversations. Could I ask you fifteen minutes of questions about how your team did it last year? I'm not selling anything." That message works, and the last sentence is doing most of the work.

Prepare a script, then be willing to leave it

Write your questions in advance. Not to read mechanically, but because unprepared interviews drift into pitching within four minutes.

A structure that works:

1. Context (2 minutes). Who they are, what their role is, what the relevant activity looks like for them. Warm-up, but it also tells you whether they are the person you think they are.

2. Last occurrence (7 minutes). "Tell me about the last time you had to do X." Then walk through it step by step. What triggered it. Who else was involved. What tools they used. Where it got stuck. How long it took. What they did when it went wrong.

This section is the heart of the interview. Stay in the past tense. Keep asking "and then what happened?"

3. Cost and workaround (3 minutes). What did that cost you — hours, money, marks, a lost opportunity? What have you tried to make it better? Why did that not work?

4. Alternatives (2 minutes). What else did you consider? What do others in your position do? Have you paid for anything to help?

5. Referrals (1 minute). Who else deals with this that I should speak to?

Notice that your idea is not in the script. If they ask what you are building, answer briefly and honestly at the very end, after the questions are done.

Ask about the past, never about the future

This is the one rule that separates useful interviews from flattering ones.

People are poor predictors of their own behaviour and generous with hypothetical enthusiasm. They are far more reliable when describing what they actually did.

Replace future questions with past ones:

Avoid	Ask instead
Would you use an app for this?	What did you use the last time?
Would you pay ₹300 a month?	What have you spent money on to solve this?
Do you think this is a good idea?	Walk me through the last time this went wrong.
How often would you use it?	How many times did this come up last month?
What features would you want?	What did you do when the tool you had didn't cover it?

The left column produces opinions. The right column produces facts.

Six mistakes that ruin an interview

Pitching. The moment you describe your solution, the conversation becomes about your feelings, and the other person switches into politeness mode.

Leading questions. "So that was frustrating, right?" tells them the answer you want. Ask "how did that go?" and wait.

Filling silence. After a question, count to five before speaking again. The most valuable sentences usually arrive in that gap.

Accepting abstractions. "It's usually a mess" is not data. Ask "tell me about a specific time."

Hearing only agreement. Write down objections and hesitations verbatim; they are more informative than praise.

Interviewing the wrong person. A person who supervises the task is not the person who does it. Talk to whoever's hands are on the problem, and separately to whoever controls the budget.

Record it properly

Two people is the ideal setup: one asks, one writes. If you are alone, ask permission to record, and take rough notes anyway.

Capture:

• Direct quotes, in their words, not your paraphrase.
• Numbers they mention: frequency, hours, rupees, headcount.
• The tools and workarounds they named.
• Anything that surprised you or contradicted your assumptions.
• Their emotional peaks — where they got animated, irritated or resigned.

Write up the notes within an hour. Memory degrades fast, and reconstructed notes tend to be reconstructed in the direction of your hopes.

Turn conversations into decisions

After five or six interviews, put every note side by side and look for repetition:

• Which problem descriptions came up unprompted more than twice?
• Which workarounds appeared repeatedly?
• Where did people disagree, and does the disagreement split by segment — year of study, role, size of business?
• Which of your original assumptions now look weaker?
• What did you hear that you had never considered?

Then make one of three explicit calls: continue with the current problem definition, narrow it to the segment where the pain was sharpest, or drop it.

Write the decision down along with the evidence. A month later, when enthusiasm has shifted, you will want the record.

Common questions from first-time interviewers

"Nobody will talk to me — I'm just a student." Being a student is an advantage in recruitment. People help students in a way they do not help salespeople. Say clearly that you are researching and not selling.

"How many is enough?" Stop when new conversations stop surprising you. That usually happens somewhere between eight and fifteen for a narrow segment.

"They said they loved it." Then you pitched. Go back to past-tense questions.

"Everyone described a different problem." Your segment is too broad. Narrow it — one year group, one role, one type of business — and interview again.

What comes next

Interviews tell you whether a problem is real, who has it most acutely, and what people already do about it. They cannot tell you whether your solution works. That requires putting something in front of someone and observing what they do with it.

The mistake is to jump from ten good interviews straight into three months of building. The step in between is much smaller: the least you can construct that lets a real person get the outcome they described wanting, so you can watch whether they actually use it.