Thinking
Keep a one-page record of important decisions
Record the problem, authority, evidence, alternatives, and accepted tradeoffs so that later reviews can examine reasoning as well as outcomes.
Preserve the problem as it existed at the time
Decisions are usually made with incomplete information and limited resources. Once an outcome is known, it is easy to import later knowledge into the earlier choice. A decision note preserves the conditions under which the judgement was actually made.
Begin with the problem, the reason a decision is needed now, and the scope it covers. The background does not need to reproduce the entire history, but it should allow a later reader to understand the task.
Keep the desired outcome separate from a preferred solution. If the problem statement already assumes one approach, comparison can become a formality rather than a genuine examination of available actions.
Clarify authority, constraints, and reversibility
Identify who makes the decision, who contributes information, and who carries it out. Participation in a discussion does not necessarily confer final authority, and responsibility for execution does not imply permission to alter every condition.
Record constraints such as time, available capability, dependencies, and essential requirements. Distinguish fixed limitations from working assumptions so that the team does not unnecessarily restrict its options.
Scale the depth of the record to the decision. A readily adjustable choice with limited impact can remain brief. A difficult-to-reverse commitment affecting several parties needs a fuller account of evidence and reconsideration conditions.
Separate facts, assumptions, and preferences
Facts describe available information and need an attributable basis and scope. Assumptions connect what remains unconfirmed and need a way to be checked. Preferences describe priorities that should be open to discussion. All three can inform a choice without being interchangeable.
When information is incomplete, identify the uncertainty and the most significant gaps. Avoid giving weak judgements an appearance of precision through unsupported numbers. Readers should be able to distinguish strong evidence from tentative interpretation.
Gathering more information also consumes time and effort. Prioritise information that could change the choice or its implementation. Material that only makes the document longer need not postpone an otherwise workable decision.
Compare actions that are genuinely available
Alternatives should be feasible at the time and assessed against the same objective. Maintaining the current approach, narrowing the scope, proceeding in stages, or waiting may be valid options when their conditions are made explicit.
Choose comparison criteria from the objective and constraints. Too many criteria can obscure the main differences, while a single attractive feature can hide substantial costs. Exclude options that fail essential conditions before comparing the remaining tradeoffs.
Do not add alternatives after the event merely to make the process look thorough. An accurate account of the real choice is more useful for learning than a polished reconstruction.
Explain the choice and the costs accepted
The rationale should connect the selected option to the evidence, constraints, and objective. Repeating its advantages is insufficient. Explain why the other serious options were less suitable under the conditions that applied.
A choice may give up flexibility, speed, completeness, or ease of maintenance. Recording those costs helps the people implementing it distinguish an accepted compromise from a deviation that requires attention.
Also identify the conditions that would weaken the original rationale. This turns uncertainty into something to monitor, with ownership and a possible response, rather than leaving it as a list of concerns.
Use new evidence to trigger reconsideration
Review can be triggered by elapsed time, completion of a milestone, or a change in a critical assumption. State who will review, what they will examine, and how a revised decision will be made.
Separate outcome quality from decision quality. A well-supported choice can encounter unexpected conditions, while a poorly supported choice can produce a favourable result. Review the reasoning available at the time, the execution, and subsequent changes together.
When revision is needed, preserve the original note and add the new decision and its basis. Replacing the earlier record removes the very information needed to understand why the direction changed.
Keep the record easy to maintain
One page is a principle of low maintenance and useful density, not a rigid physical limit. The main record can hold the problem, options, evidence, tradeoffs, owner, and review conditions, with supporting material linked separately.
Store the note where relevant collaborators can find it, with a clear date and status. Connect significant decisions to their resulting actions so that the record remains part of the work rather than a detached archive.
Over time, these notes help a team explain its judgement, recognise recurring assumptions, and respond to changing conditions. Brevity, accuracy, and retrievability make the practice easier to sustain than turning every choice into a formal report.