Track assumptions in a five-field register: claim, impact if false, evidence, status, next test
Maintain an assumption register with five components for each assumption: the specific testable claim, what would change if false, current evidence for/against it, validation status, and next action to test it—reviewing weekly for active projects.
Why This Is a Rule
Every plan rests on assumptions, and untested assumptions are the primary source of plan failure (see Extract and classify every assumption before committing — test high-importance, high-vulnerability first). But extracting assumptions is only the first step — managing them through the project lifecycle requires structure. The five-field register provides that structure, tracking each assumption from identification through testing to validation.
Testable claim: the assumption stated precisely enough to be falsified. Not "customers want this" but "at least 30% of surveyed customers will rank this feature in their top 3." Impact if false: what changes in the plan if this assumption is wrong? This determines priority — high-impact assumptions get tested first. Current evidence: what supports and contradicts this assumption right now? Validation status: untested, partially tested, validated, falsified. Next action: the specific step to test or validate this assumption.
Weekly review keeps the register active. Without review, it becomes a static document created during planning and never consulted — assumptions accumulate without testing, and the register provides no value.
When This Fires
- During project planning when assumptions are being extracted (Extract and classify every assumption before committing — test high-importance, high-vulnerability first)
- Weekly during active projects to track assumption validation progress
- After any event that might have validated or falsified a registered assumption
- When a project outcome surprises you — check which assumptions were wrong
Common Failure Mode
Creating the register during planning and never reviewing it. The register's value comes from weekly interaction: checking validation status, updating evidence, executing next actions. A register created once and filed away is a planning artifact, not a management tool.
The Protocol
For active projects: (1) Extract assumptions into the register with all five fields populated. (2) Prioritize by impact-if-false: which assumptions would kill the plan if wrong? (3) Weekly: review each assumption. Update evidence. Execute next-action tests for the highest-priority untested assumptions. Update validation status. (4) When an assumption is falsified → immediately evaluate plan impact and adjust. (5) When an assumption is validated → reduce monitoring frequency. The register should shrink over project life as assumptions move from untested to validated/falsified.