Customer discovery interviews are only useful if they surface the truth. Here is how to ask the right questions, listen well, and come away with signal that actually shapes what gets built.
Startup customer interviews are one of the most valuable things a founder can do early on. They are also, surprisingly easy to do in a way that feels productive but produces almost nothing useful.
The core problem is this: people are kind. When someone shares an idea with genuine excitement, most people will find something encouraging to say about it. In India especially, where hospitality and warmth are deeply embedded in how conversations work, the instinct to be supportive is strong. A founder can walk away from twenty conversations feeling thoroughly validated — and still be building something the market does not actually want.
This is not anyone's fault. It is just what happens when discovery interviews are structured around pitching and seeking approval rather than learning. The good news is that a few changes in approach — to the questions asked, the listening that happens, and the way conversations are set up — can change everything about what comes out of them.
The shift worth making before any interview: the goal is not to find out whether people like the idea. It is to find out whether the problem is real, how it actually shows up in their lives, and what they are already doing about it. The idea itself does not need to come up at all.
There is a pattern that comes up again and again in early-stage discovery. A founder describes the problem they are solving. The person across the table recognises it, agrees it is a real issue, says "yes, I would definitely use something like that" — and the founder leaves feeling confirmed. The problem is that hypothetical enthusiasm and actual behaviour are very different things.
The research on this is consistent: people are poor predictors of their own future behaviour, especially for things that do not yet exist. "Would you use this?" is almost never a useful question, because the honest answer to something hypothetical, asked by someone who clearly cares about the answer, is almost always yes.
What is useful is the past. What people have actually done — the workarounds they have built, the money they have already spent, the time they have already lost — is a far more reliable signal than anything they say about what they might do in the future.
The single most important shift in how to frame discovery questions is moving from future hypotheticals to past experiences. "Would you use this?" is a weak question. "Tell me about the last time this came up for you" is a strong one.
Past behaviour is specific, anchored in reality, and harder to dress up. When someone describes an actual incident — how it happened, what they did, how long it took, how frustrated they were — that is the kind of signal worth building on.
The moment an idea is introduced, the conversation shifts. The other person starts thinking about the idea rather than their experience. They begin to evaluate rather than recall. And because most people are naturally supportive, their feedback skews positive in ways that are hard to calibrate.
The most useful discovery interviews are ones where the founder barely speaks about what they are building. The conversation is entirely about the other person's world — their problems, their processes, their frustrations, their current solutions. The idea can come up at the very end as a natural coda, but it should not be the reason for the conversation.
Facts tell part of the story. Emotion tells the rest. When someone describes a problem with visible frustration, with a sigh, with a phrase like "it is such a headache" or "honestly, it drives me crazy" — those moments deserve to be followed. That is where the real signal lives.
A question as simple as "that sounds frustrating — can you say a bit more about that?" can open up a much richer conversation than moving on to the next prepared question. Good discovery interviewers are more curious than they are systematic.
Every meaningful problem already has some kind of solution in use — even if that solution is a spreadsheet, a WhatsApp group, or sheer manual effort. Understanding the current solution is one of the most revealing parts of any discovery conversation, because it shows how seriously the problem is taken and what someone is already willing to invest in addressing it.
If someone has built an elaborate workaround for a problem, that is strong signal. If they have not thought about it much and are not currently doing anything to address it, that is also signal — and it suggests the problem may not be as painful as it initially appeared.
A problem that costs people real time, real money, or real peace of mind is a problem worth solving. A problem that is merely inconvenient is much harder to build a business around. Discovery interviews are a good place to understand which category a problem falls into.
The most useful way to do this is not to ask directly ("how much would you pay?") but to understand the actual cost of the current situation. How long does the current workaround take? What happens when it goes wrong? Has anyone lost a client, missed a deadline, or had a difficult conversation because of this problem? Those are the moments that reveal whether the pain is real.
The quality of a discovery conversation depends entirely on who is in it. Talking to someone who is adjacent to the problem — who has heard about it, who knows someone who faces it, who faces a softer version of it — produces very different insight from talking to someone who lives with it daily.
In India, this often means being more deliberate about who is actually in the room. It means going beyond alumni networks and warm introductions to find people who truly fit the customer profile. It means choosing specificity — "logistics managers at mid-size FMCG distributors in Tier 2 cities" — over convenience. The further from the actual customer the conversation is, the less reliable the signal.
The notes taken after a discovery interview are often where the most important distortions happen. Memory is selective, and it tends to select for the things that confirm existing beliefs. A positive comment gets remembered; a hesitation gets forgotten.
The most useful notes are verbatim quotes — the exact words someone used to describe their experience. "It's a bit annoying" and "it's a complete nightmare" are very different things, and the difference matters. Keeping actual quotes, rather than paraphrased summaries, makes it much easier to look back later with fresh eyes and see what was actually said.
Twenty interviews produce a lot of raw material. The work of making sense of it is as important as the interviews themselves.
A simple approach that tends to work well: after each conversation, write down the three most important things heard — not conclusions, but actual moments from the conversation. Then, after a batch of five to ten interviews, look across all of them for patterns. What phrases came up repeatedly? What frustrations appeared again and again? What workarounds were most common? Where did the energy in the room change?
Patterns across multiple conversations are what transform individual interviews into real insight. A single person saying something is interesting. Five people saying a version of the same thing independently is a signal worth acting on.
It is equally worth paying attention to what did not come up. If a problem that seemed central to the idea barely registered in any conversation, that is worth sitting with — not dismissing, but genuinely exploring.
A useful rule of thumb is to keep going until the conversations start feeling repetitive — until the same themes, the same frustrations, and the same workarounds are coming up without much new being added. For most early-stage ideas, that point tends to arrive somewhere between 15 and 30 conversations. For very niche B2B problems, even 8 to 10 deep conversations with the right people can be enough to find clear patterns.
Warm introductions are the most reliable route — through alumni networks, professional communities, LinkedIn connections, or accelerator contacts. Cold outreach also works surprisingly often, especially when the message is honest and specific: "I am exploring a problem in [space] and would love 20 minutes to hear about your experience — not selling anything, just learning." Most people are willing to share their experience when the ask is genuine and low-stakes.
That is the discovery process working as it should. The point of these conversations is to learn — and sometimes what is learned is that the initial framing of the problem was off, the customer segment needs refining, or a different version of the idea is actually more interesting. Finding that out through conversations is far less costly than finding it out after months of building.
The Karak is a space where early-stage founders in India share what they are learning — including from customer conversations that went in unexpected directions. If this is something being worked through right now, it is a good place to compare notes.
Visit The Karak →Written by the team at The Karak — a space for early-stage founders in India to share real experiences, honest reflections, and the kinds of conversations that are hard to find elsewhere.
