Falsifiable Hypotheses: What Week 0 Produces

A hypothesis that cannot fail cannot be supported either. Before research starts, a Sprint fixes the hypothesis and the pass/fail criteria so no result can be reinterpreted afterward.

By Harinderpal Hanspal on December 2025. Updated October 2026

"Manufacturers want better predictive maintenance" cannot fail. It is vague enough to be true of nearly every manufacturer at some price, eventually, so no interview result can refute it and none can meaningfully support it.

Four artifacts are signed off before research begins, so the criteria cannot move afterward A boxed list of four Week 0 artifacts: a hypothesis that can fail, pass and fail criteria, an ideal customer profile, and an access path. An arrow passes through a red gate marked criteria cannot move, into interviews, which lead to a verdict of Proceed, Pivot, Reset or Stop. WEEK 0, SIGNED OFF FIRST 1 Hypothesis that can fail 2 Pass and fail criteria 3 Ideal customer profile 4 Access path criteria cannot move Interviews scored against them Verdict Proceed, Pivot, Reset or Stop
Four artifacts are signed off before research begins, so the criteria cannot move afterward

A falsifiable version names a persona, a segment, a price and a budget path: plant managers at discrete manufacturers with three or more sites will pay a stated annual contract value for this capability, approved through the operations budget. It can fail on any of the four, and that is what makes it testable. With 98% of manufacturers exploring or considering AI-driven automation, per Redwood Software's 2026 survey, only this version escapes the polite agreement the generic one meets everywhere.

Before research begins, a Sprint produces four signed-off artifacts, which the method calls Week 0:

  1. The hypothesis, written in falsifiable form.
  2. The pass and fail criteria: the evidence thresholds that count as confirmation or refutation.
  3. A definition of the ideal customer profile precise enough to build a target account list.
  4. The access path: whether the named accounts come from the client's network or from cold sourcing.

The criteria come first. Success defined after the fact always matches what was found. If the criteria can move after the interviews, the Sprint becomes an expensive way to confirm a decision already made. Declaring them first is what lets the verdict survive scrutiny from a board or an investor.

Go deeper: Too successful to stop, too unproven to scale, Test AI-written code by what a silent failure would cost and the Sprint page shows the six phases that follow it.