discovery-experimentslisted

# Discovery experiments ## Core principle An experiment exists to change a belief. Before designing one, write the threshold that would change yours, and write it before you see any data. A test whose success criterion is decided afterwards is not a test, it is a search for a number that supports the plan. Second principle: **test the riskiest assumption with the cheapest method that would actually move your belief.** Not the easiest assumption. Not the most impressive method. ## The sequence ``` solution → assumptions → map by importance x evidence → pick top-right → choose cheapest sufficient test → set threshold BEFORE running → run → read honestly → record what changed ``` Skipping the mapping step is why teams test the thing that is easy to test. ## Step 1: surface the assumptions For any solution, list every statement that must be true for it to work. Five categories (Cagan's four risks plus Torres's ethical): | Category | The question | |---|---| | **Desirability / value** | Do they want it enough to change behaviour? | | **Usability** | Can they work out how to use it? | | **Feasibility** | Can we build it, at acceptable cost and performance? | | **Viability** | Does the business work: cost to serve, price, channel, legal, support? | | **Ethical** | Could this harm someone, and would we defend it in public? | **Write each as a falsifiable statement, not a topic.** "Pricing" is a topic. "Merchants will pay 5 percent of order