How to run your first 10 customer interviews without leading them

Why the first ten interviews decide everything

Most founders skip discovery interviews because they feel slow compared to building. In practice ten conversations take about a week and remove the single biggest risk in a new product: that nobody has the problem badly enough to change what they do today.

The goal is not enthusiasm. Enthusiasm is free and means nothing. The goal is evidence about behaviour that already happened — money spent, workarounds built, tools abandoned, hours lost. That is the raw material of idea validation, and it only shows up when you ask about the past.

Recruit for recency, not for job title

The cheapest filter is recency. Instead of looking for a persona, look for someone who hit the problem in the last 30 days. Recent pain means the details are still accurate and the workaround is still visible.

• Write one screening sentence: 'Have you tried to do X in the last month?'
• Ask for a referral at the end of every call — the second wave costs nothing.
• Cap yourself at three people who already like you; friends distort the signal.
• Keep a recruiting log with source, date, and whether they qualified.

Write the questions before you write the product

A good interview script is short: five to eight open questions, all backward-looking, none of them mentioning your solution. Anything that starts with 'Would you' invites a polite lie. Anything that starts with 'Tell me about the last time' produces a story you can verify.

If you want a starting script, generate one from your own idea instead of copying a generic list — the interview question generator produces non-leading customer discovery interview questions in about a minute and you can edit them before your first call.

Run the call: 80% listening, verbatim notes

Spend the first two minutes setting expectations: you are researching, you are not selling, and there is no wrong answer. Then shut up. The most valuable moments come from silence after an answer, not from your next question.

Write down what people actually said, in their words. Paraphrasing is where evidence quietly turns into confirmation bias. Verbatim quotes are also what convinces a co-founder or an investor later.

• Record when allowed, and always ask permission on the recording.
• Note the workaround, the tool it replaced, and what it costs today.
• Ask what they tried and abandoned — abandonment is a stronger signal than intent.
• Close with: 'Who else should I talk to about this?'

Score every interview against one written assumption

Before each call, write the single risky assumption the call is meant to test — for example, 'operations leads at 20-50 person agencies already pay someone to reconcile invoices manually'. After the call, mark it supported, contradicted, or unclear, and paste one quote as the reason.

Ten scored interviews give you a pattern instead of a feeling. If seven contradict the assumption, you have saved yourself six months. If seven support it, you have earned the right to run a paid pilot.

Know when to stop

Stop interviewing on a given assumption when three consecutive calls tell you the same thing. Saturation, not a round number, is the real signal. Then move to the next riskiest assumption and repeat with a fresh script.