Outcomes
Four outcomes, and what each one does not mean
A Foundry review returns one of four startup evidence outcomes: Build, Iterate, Kill or Abstain. Each applies only to the single assumption reviewed, never to the company or the founder. Iterate and Abstain are real answers, not failures, and none of the four predicts whether a startup will succeed.
Iterate
Not yet. There is signal here, but not enough to build on, and not enough to stop.
Either a comparable-or-stronger alternative exists and is worth pursuing instead, or real uncertainty remains that more evidence could resolve.
Returned when the evidence is usable but something specific is unresolved, or when your assumption has weakened and an alternative has comparable or stronger evidence. The result lists exactly what is unresolved and the next evidence step.
- Not: The idea failed.
- Not: Nothing was learned.
- Not: This is a stall or a non-answer.
Abstain
A responsible evaluation isn't possible yet.
The decision statement or the evidence itself is not usable as submitted -- so no BUILD, ITERATE, or KILL call would be honest. This is a deliberate, correct result, not an error.
Returned when the assumption is not clearly stated, the question already assumes its answer, or the evidence cannot be used as submitted. Foundry says which, instead of guessing.
- Not: The system is broken.
- Not: A soft KILL.
- Not: A soft BUILD.
Build
Current evidence justifies a bounded next step.
The evidence gathered for this specific assumption clears the required bar for its risk level, so a defined next build step is justified now.
Returned when the evidence meets every requirement for the decision risk you chose, including an attempt to disconfirm it, and at high risk your explicit confirmation.
- Not: Product-market fit is achieved.
- Not: The startup will succeed.
- Not: No further evidence is needed.
Kill
This assumption, as tested, should stop.
Every required safeguard (a pre-registered falsifier, a real repair attempt, repeated contradiction over time, independent evidence, and your explicit confirmation) was satisfied, so this assumption in its current form should stop.
Returned only when every safeguard is met: a falsifier written before the evidence, a real repair attempt, contradiction repeated at least seven days apart, independent evidence from two channels, and your confirmation.
- Not: Stop the company.
- Not: Quit entrepreneurship.
- Not: The venture will fail.
How the policy chooses between them
The policy checks in a fixed order and stops at the first outcome that applies: Abstain, then Kill, then Iterate towards a better-supported alternative, then Build, and otherwise Iterate on the open uncertainty. The marks above are one ink and four shapes on purpose. An outcome is not a grade, and none of them is coloured as good or bad.
If an outcome would apply but needs your confirmation, the result says it was blocked and why, and continues down the order. It never returns Kill, or a high-risk Build, without that confirmation.