Editorial draft available to read. Factual and language review is pending. Examples are illustrative.
Separate observation and assumption
Label inputs measured, estimated or unknown, with source, date, owner and unit. Repetition does not turn opinion into observation. For estimated time savings, specify included tasks and the review work that remains after implementation.
Show sensitivity
Build coherent scenarios by varying influential assumptions. Include implementation, adoption, operations and exceptions. Explain what must happen for each scenario; it is not a guaranteed forecast. Avoid decimal precision when inputs are broad, unconfirmed ranges.
Assign the next check
Hypothetical example: report preparation is estimated at two hours weekly but unmeasured. Observe several runs and separate preparation from interpretation before annualising savings. Each important assumption needs evidence or a pending test and an explanation of how an incorrect assumption changes the decision.
Key takeaways
- Expose the data, estimates and conditions behind the decision.