Concept
Customer conversations
Customer conversations are the founder's cheapest experiment: structured, repeated talks with real people who have the problem, run to extract facts about their past behaviour rather than opinions about your product.
Fitzpatrick's larger argument is that conversations are not a phase before building but a practice that never stops, and that most founders do them badly for one reason: they are looking for approval. A conversation run for approval produces compliments, fluff and hypotheticals, all of which feel like progress and none of which are. A conversation run for facts produces three things. Facts about the problem: how often it happens, what it costs, what the workaround is. Emotional signals: the moment the person leans forward or goes quiet tells you where the pain actually lives. And commitments: whether they will give you something real, which is the only proof that the problem matters enough. Keep the conversations casual and short, take notes on paper, and share what you learned with the team the same day; a learning that stays in one founder's head is a learning the company does not have.
A real-life example
Two founders in Bursa selling software to logistics firms agreed that neither would build a feature the other had not heard described by a customer, in the past tense. Their notes went into a shared page every evening. Within three weeks the page showed the same complaint from eight dispatchers, about a step the founders had not known existed, and that step became the product.
How to use it
- 1Schedule conversations as a standing habit, three to five a week, not as a research phase that ends.
- 2Write down facts, emotions and commitments separately after each conversation; a page of compliments is a page of nothing.
- 3Share the notes with your co-founders the same day, and let a feature into the build only when a customer has described the need in the past tense.