Getting Out of the Building

The central claim

At the start of a new venture, everything is a guess: who the customer is, what problem they have, what they will pay, how they will be reached. These guesses feel like knowledge because they are held confidently, but no amount of internal discussion converts a guess into a fact.

The only source of facts is outside — talking to people who might buy, watching what they do, and running experiments. Every hour spent refining a plan internally instead of testing it externally is an hour spent making a guess more detailed rather than more likely to be true.

What discovery replaces

The traditional sequence — write a plan, build the product, launch, then find out — concentrates all the learning after all the spending. Discovery inverts it: learn first, cheaply, and build only what the learning justifies.

This is not an argument against planning. It is an argument for treating the plan as a set of hypotheses with an explicit programme for testing them, rather than as a description of the future.

Problem before solution

The most common discovery failure is starting with the solution. A founder in love with their idea asks questions designed to confirm it, hears encouragement, and builds something nobody needed. The order matters:

1. Is the problem real, frequent and painful?
2. Who exactly has it?
3. What do they do about it now, and what does that cost?
4. Only then: would this particular solution be better?

The founder's two biases

• Confirmation bias. You will unconsciously seek evidence supporting your idea and discount evidence against it. The defence is to state in advance what would prove you wrong, and to look for it deliberately.
• The curse of knowledge. You understand your solution deeply, so explanations that seem obvious to you are incomprehensible to others. Confusion in an interview is data about your communication, not about the listener.

What good discovery looks like

• The founder does it personally. Delegating early customer conversations to a researcher removes the person who most needs to hear the answers.
• It is continuous rather than a phase that ends.
• It produces uncomfortable findings, which is the sign it is working. Discovery that only confirms what you believed was not discovery.
• It changes what gets built.