Run ideas through a kill test

A founder’s job is not to protect ideas. It is to expose them to reality while the cost of being wrong is still small.

Most idea lists are collections of unpriced assumptions. Convert each idea into a testable thesis:

We believe [specific customer] will pay [price or economic commitment] for [result] because [observed problem] currently costs [time/money/risk]. We can test this by [date] for no more than [budget/hours].

The twelve gates

An idea advances only if you can answer:

1. Who experiences the problem? 2. Who controls the budget?

3. What happened the last three times? 4. What is the measurable cost of inaction? 5. What solution or workaround exists now? 6. Why is the current option inadequate? 7. Why can you reach and serve this buyer?

8. What is the smallest paid proof? 9. What must be true for delivery to be safe and legal? 10. What would falsify the thesis? 11. What is the maximum affordable loss? 12. What exact evidence triggers continue, change, or stop?

If several answers are “I assume,” the next step is not development. It is research.

Pre-commit the stop rules

Before emotion and sunk cost grow, write stop conditions. Example:

• • after 30 qualified conversations, fewer than five describe the problem as urgent;
• • fewer than three will introduce the economic buyer;
• • no one accepts a paid pilot after ten relevant demonstrations;
• • delivery requires access, accuracy, or liability the team cannot responsibly manage;
• • the sales cycle and acquisition cost make the present model uneconomic;
• • the founder repeatedly avoids this customer and domain after real exposure.

Stop rules do not mean every target must be hit immediately. They force a deliberate review rather than silent drift.

The evidence ladder for ideas

· GE ch. 10; ASC pp. 50–56; TABLE pp. 72–74. Mechanisms paraphrased; judgment and examples are labelled.