The work before stage one is a sequence of decisions: what is true, what must be true, and what evidence would justify committing resources.
Ideas are inputs, not progress
A room can produce hundreds of ideas and still avoid the decision that matters. Volume feels productive because it creates artifacts. Stage-zero progress is different: uncertainty is reduced only when the team names an assumption, gathers evidence, or makes a reversible choice.
Divergence matters, but it should open model options rather than generate disconnected features. Each option must say who changes behavior, what value is exchanged, how the offer reaches them, and why the system can sustain itself.
Move from facts to decisions deliberately
Start with what is directly observed. Mark interpretations separately. Turn the most consequential interpretations into assumptions, then rank them by uncertainty and damage if wrong.
The next experiment should target the assumption that can most change the model. A good test has a representative participant, a real behavior or costly commitment, an observable signal, and a decision rule agreed before results arrive.
Stage zero ends when the team can make a decision without pretending uncertainty has disappeared.
00z.ai
Stage one is an earned commitment
The purpose of stage zero is not to remove all risk. It is to know which risks remain, why they are acceptable, and what execution will learn next. A team is ready when it can explain the model, evidence, economics, advantage, and explicit reasons to proceed.
That is the handoff from exploration to ownership. The business model becomes a shared operating thesis rather than a founder story or workshop output.
