Decisions
What was decided, why, and what changed. Your future self will thank you.
When to Log a Decision
Not everything is a decision. Here is the line.
What IS a decision:
- A project pivot or direction change
- Adopting or dropping a tool
- Changing your approach to a problem
- A strategy shift that affects multiple tasks
- A pricing change or business model adjustment
What is NOT a decision:
- Variable naming or code formatting choices
- Routine tasks that follow existing processes
- Daily tweaks and minor adjustments
What Makes a Good Decision
A decision without reasoning is just a note. Every decision file must include these fields.
| Field | Description |
|---|---|
| Title | Clear, specific name for the decision. Not "Decided stuff about pricing." |
| Context | What situation triggered this decision. What problem needed solving. |
| Options considered | At least 2 alternatives. If you only considered one option, you did not make a decision. |
| Reasoning | Why you chose this option over the others. Not just what you picked, but why. |
| Consequences | What changes as a result. Which tasks, projects, or workflows are affected. |
| Status | Current state: proposed, decided, implemented, or reversed. |
The key requirement: options AND reasoning. A decision with only one option is not a decision. A decision without reasoning is a coin flip.
Decision Lifecycle
Decisions move through four stages:
- Proposed. Options laid out, reasoning drafted, waiting for final call.
- Decided. Choice made, reasoning locked. This is the source of truth.
- Implemented. The decision has been acted on. Tasks created, changes shipped.
- Reversed. Rare but it happens. Only with new evidence and documented reasoning.
Reopening Rule
Settled questions stay settled. A decision can only be reopened with:
- New evidence that was not available when the decision was made
- Changed constraints that invalidate the original reasoning
- Failure signals from the implementation that prove the approach does not work
This prevents decision churn. Without this rule, you end up revisiting the same questions every week and never building momentum. If you want to change direction, write a new decision that references the old one.