Decisions ยท 7 minute guide
How to keep a decision journal without overthinking it
Capture the minimum information needed to improve judgment: options, assumptions, confidence, evidence and a review date.
The short answer
Record only decisions with meaningful stakes or learning value. Before acting, write the choice, alternatives, key assumption, expected result, confidence and review date. At review, score the reasoning separately from the outcome.
Good outcomes can come from bad decisions
If you judge a decision only by its outcome, luck becomes invisible. A risky bet can work once, while a sound decision can fail because of an unlikely event.
A decision journal freezes what you knew and believed before the result. This makes hindsight bias visible and gives you material for improving judgment.
The six-line decision record
- 1. Decision State the choice and deadline in one sentence.
- 2. Alternatives List the realistic options you rejected, including doing nothing.
- 3. Key assumption Name the belief that most needs to be true for the choice to work.
- 4. Expected result Describe an observable result and the period in which it should appear.
- 5. Confidence Give a probability, not a label such as likely. Use 60%, 75% or another explicit estimate.
- 6. Review Set the earliest date when enough evidence should exist to continue, change or stop.
Review reasoning and outcome separately
First ask whether the process used the best available information, considered alternatives and matched the downside. Then ask whether the expected result occurred.
Four combinations are possible: good reasoning and good outcome, good reasoning and bad outcome, bad reasoning and good outcome, or bad reasoning and bad outcome. The third is the most dangerous because success can reinforce a poor process.
Questions
Should I record every decision?
No. Record choices with meaningful downside, repeated importance, uncertainty or learning value.
How precise should confidence be?
A rough probability is enough. Consistent estimates become useful when you compare them with outcomes over time.
When should the review happen?
At the first date when the decision should have produced enough evidence to change course.