I did 60+ customer interviews before I had a product. It was the wrong order.
Everyone says "talk to your customers."
Good advice. Also the advice most commonly used to avoid shipping.
I did 60+ customer discovery calls in the early days of Genie 007. I knew more about my potential customers' pain than they did. Patterns, themes, full Notion databases of insights. Proper product discovery, done right.
What I didn't have: a working product anyone could actually use.
The uncomfortable thing nobody tells you: customer interviews tell you what people SAY they want. Revenue tells you what they'll ACTUALLY pay for. These are completely different data sets.
At some point, talking more is a displacement activity. Useful displacement. But still displacement.
The moment I realised this was about 4 months in. I was scheduling yet another discovery call and noticed I was relieved. Not to learn more. Because it postponed the judgment of putting something real in front of people and seeing them not care.
I was using "more research" as protection.
Once I stopped hiding behind interviews and shipped the thing, I learned more in two weeks than in three months of calls.
Here's the distinction that changed everything for me:
Customer discovery is for understanding the PROBLEM. But understanding the problem doesn't mean you understand your SOLUTION yet. These require completely different methods.
You understand the problem through talking. You understand whether your solution works by watching people actually try to use it. Watching, not asking.
When you ask someone "would you pay for X?", they imagine the ideal version of X. When they try your actual product and stop halfway through, that's the real answer.
The full loop:
- Talk to understand the problem
- Build something real (not perfect)
- Watch people use it without you explaining it
- Do 10 more conversations AFTER they've used it
Most founders stop at step 1 or jump to step 2 and skip the watching entirely.
The founders I see stuck in discovery mode for 6+ months usually have a fear problem, not a research gap.
The research has diminishing returns past a certain point. What you're missing isn't more interviews. It's the courage to put an imperfect thing in front of someone and watch what actually happens.
That's the real test.
Has anyone else used "customer research" as a delay tactic? What snapped you out of it?
The fact that you felt relieved when scheduling yet another interview may be the most important signal in the whole post. At that moment, the interview was no longer only a learning tool. It had also become a protective layer that delayed exposure to real feedback.
There is an old Chinese idea that fits this closely: “the unity of knowledge and action.” It does not simply mean “think less, do more.” It means action is not a second stage that begins after knowledge is complete; action is part of how knowledge becomes complete. Interviews tell you how users describe the problem, but only when they actually use, abandon, or pay for something do you know whether your understanding of the solution holds.
I would put one gate in front of the next interview: before booking it, write, “If the answer is X, which specific product action or experiment will I change?” If you cannot answer that, do not book it yet, because the conversation will probably add information without adding understanding. If you can answer it, turn the result into the smallest observable action so behavioral evidence follows verbal evidence.
So the problem may not be that 60 interviews were too many. It may be that the loop between knowing and doing had broken. One small experiment after three interviews may get you closer to real knowledge than thirty uninterrupted interviews.
“Knowledge is the beginning of action, and action is the completion of knowledge.” — Wang Yangming, Instructions for Practical Living. In plain terms: knowledge should lead into action, and action also completes what you know; the two cannot be separated into unrelated stages.
The Wang Yangming framing is the one that actually lands.
I'd been telling myself the interviews were about clarity. But when I felt relief at booking number 61, that was the tell. Relief isn't a learning signal. It's an avoidance signal.
Your gate would have saved me probably 40 calls. Before booking, write down which specific action changes if the answer is X. If you can't write that down, the interview is filling an anxiety gap, not an information gap.
Since shipping rough versions of Genie 007 instead of interviewing more: the conversations got 10x more useful. People were reacting to something real, not describing hypothetical preferences. The gap between "yes that sounds useful" and "I've used this 5 times this week" is where all the real signal lives.
You put it very clearly: relief is not a learning signal, but an avoidance signal. More importantly, you have already tested that insight through what happened after you shipped. Conversations around a real product became far more useful than interviews about a hypothetical one.
That means this gate is no longer merely my suggestion. It is something you have demonstrated through action. Thank you for sharing what happened afterward. It makes me even more convinced that Wang Yangming's “unity of knowledge and action” is not just a philosophical idea here, but a very practical way of working.
That framing really stuck with me writing this. Relief as an avoidance signal vs a learning signal is something I kept coming back to. You are right that the post-ship conversations were where the real learning happened. I just did not have the framework to see it until I had something real to react to.