Talking to customers to validate an idea means running 30-minute discovery interviews with people who live the problem. Ask the same questions every time. Speak for under 15% of the call. A start-a-business app like Draper takes the idea all the way to live ads; it still cannot make the calls.
This is the least glamorous work in starting a business, and it is the work most first-time founders skip. It is slow, it is repetitive, and it cannot be delegated to a model. What follows is the method Draper's founding team used during its own discovery phase.
How do you talk to customers to validate an idea?
Identify people living the problem, send a personalised outreach message, then run 30-minute interviews in focussed blocks. Ask the same questions every time, speak for no more than 15% of the call, and follow up until the frustration underneath the answer is clear.
The full workflow:
- Write down who you think lives the problem. Be specific about the situation they are in, not the demographic they belong to.
- List the people you can already reach. Start with your immediate circles, then put your network to work for introductions.
- Send a personalised outreach message using the four-part formula in the next section. No mass sends.
- Book the interviews in focussed blocks — several in a week rather than one a fortnight, so the feedback loop closes fast.
- Fix your question set before the first call and change nothing between calls.
- Run the call to 30 minutes, speaking for under 15% of it, following up until each answer bottoms out.
- Thank them and ask if they want to stay engaged. Most say yes.
- Circle back with a prototype and let them react to something real.
Why does conversation crystallise an idea faster than research?
Conversation is the best way to crystallise an idea because it forces you to say the thing out loud to someone with no reason to be kind. Every time you describe what you are exploring you collect feedback, clues and insights — plus a physical reaction from your own body that desk research never produces.
Everyone who has pitched an idea has said a sentence that made their own skin crawl. That reaction is data. It tells you which part of the idea you do not believe yet.
Repeated across dozens of conversations, this builds four things at once: a deeper understanding of the problem, a shared language for describing it, a data set you can actually compare across people, and confidence in yourself. The last one matters more than founders expect. Confidence built from dozens of conversations survives a bad month; confidence built from a spreadsheet does not.
Who should you be talking to first?
Talk to people you believe are living the problem. That phase has a name — discovery. Start in the everyday moments already around you, then work outwards through your network. Draper's own discovery for the start-a-business app ran past 150 conversations before the picture held steady.
You do not need a recruitment budget to begin. The first 10 conversations are almost always people you already know, or one introduction away. The point of those 10 is not statistical weight — it is to sharpen the questions before you spend social capital on strangers.
Be honest about who you are actually reaching, though. Friends of friends skew towards your own world, and a market that only exists inside your network is not yet a market. Our guide on whether your target market is actually real covers how to test that boundary.
How should you write the outreach message?
The discovery conversation begins before the call. Your outreach has to do two jobs: convince someone to give you half an hour, and lay the groundwork so the interview runs smoothly. A thoughtful, personalised message does both. A copy-pasted one does neither, and it lowers the quality of the answers you get.
The formula Draper used has four parts, in order:
- This is who we are.
- This is what we are exploring.
- This is why we want to speak with you.
- This is when we are available.
Part three is the one people drop, and it is the one that earns the reply. It tells the person you chose them rather than blasted them, and it primes them to arrive with the right experience in mind. Keep the whole message short enough to read on a phone.
How much should you talk during the interview?
No more than 15% of the call. Thank them, introduce yourself, say what you hope to achieve, then get out of the way. Your goal is to get the person to share what frustrated them most, and to understand the problem itself rather than test your solution.
Make the interviewee feel important, because they are. They are giving you the only unpurchasable input in early-stage business building.
Ask the same or very similar questions in every interview. Vary them and the data cannot be pieced together afterwards — you end up with 30 anecdotes instead of one data set. Ask one curious question at a time, never two or three stacked together, which will be your impulse when the silence gets uncomfortable. Listen for what they are saying and what they are not saying.
What should you never do in a discovery interview?
Never over-expose what you are building. Describing your solution poisons the interviewee's mind — from that moment they react to your product instead of describing their life, and politeness does the rest. You lose the interview and you do not find out until you launch.
Two more rules hold the method together.
Never be afraid of the follow-up. The first answer is usually the socially acceptable version. Ask again, and again, until the insight is genuinely extracted — "what made that frustrating?" is a question you can ask four times in a row.
And never run over. 30 minutes maximum, finished on time, every time. Respecting the boundary you set is what makes the person willing to take your second call — and the second call is where the prototype feedback happens.
What happens after the 30 minutes?
The relationship does not end when the call does. Thank them, then ask whether they would like to stay engaged as you learn more. 9 times out of 10 they say yes. That single question converts an interviewee into a design partner, and design partners become the people who buy first.
Circling back with a prototype does two things at once. You gather sharper feedback, because reacting to something real is easier than answering a hypothetical. And you give the person buy-in — they can see their own words in what you built.
By the time the product is ready, they buy it, because they feel like they helped design it. That is the quiet mechanism behind most early revenue, and it is why discovery interviews and finding your first 100 customers are the same activity separated by a few months.
Is talking to customers really this arduous?
Yes. There is no version of this that takes an afternoon. Identifying people, writing individual messages, chasing replies, running the calls, then writing up notes consistently enough to compare them — it is weeks of unglamorous work, and most of it happens before you have anything to show.
That is precisely why it works as validation. The cost is the signal. Anyone can generate a market report; very few founders will sit through dozens of half-hour calls where strangers describe why their idea is less interesting than they hoped.
The honest trade-off: you will spend a month on this and some of it will be wasted. Interviews will miss, people will no-show, whole segments will turn out to be wrong. Budget for that rather than treating it as failure. The wasted calls are how you find the boundary of the real market.
Where does a start-a-business app fit around this?
A start-a-business app like Draper runs the full sequence — validating the idea, building the brand, designing the product, creating the website and running live ads that put the offer in front of strangers. What it cannot do is sit on the call. The 30-minute discovery interview stays a job only a human can do.
That is worth naming plainly, and it is different from the fear that AI research is unreliable. Draper's ads prove strangers will click on the offer — real market contact, not a guess. A click is still not a payment. Converting that traffic into revenue, and hearing a customer describe the problem in their own words, stays the founder's work.
Used in the right order, the two compound. Research narrows the field so your first 10 interviews are aimed at plausible people rather than random ones, and the interviews then correct the research. Our guide on validating demand before you build covers the demand test that follows.
Want to research your idea before you build? Try Draper free →
