After a decision turns out well, it is easy to remember being more certain than you were. After it goes badly, the warning signs can seem more obvious than they felt at the time. A short decision journal gives a later review something concrete to examine.
The aim is not to record every ordinary choice. Start with a decision that involves uncertainty and will be worth revisiting: choosing a launch window, changing a service process or assigning a limited budget.
A practical entry to try
Write the decision date, the choice being made and the realistic alternatives. Then record the information available now, the assumptions you are relying on and what remains unknown. Keep facts and estimates visibly separate.
Add what you expect to happen and when you will review it. An expectation such as “the new intake form will reduce incomplete requests during the next month” is easier to examine than “this will improve everything”. Identify what you would measure before the outcome is known.
For an illustrative team decision, a manager might choose a small pilot rather than an immediate full rollout because support capacity is uncertain. The journal should preserve that reason, including the disadvantages accepted by delaying wider access. It need not pretend that the alternative was foolish.
Review the reasoning as well as the result
At the agreed date, add a new entry. Do not rewrite the original explanation to fit what happened. Ask which assumptions held, which changed and which information could reasonably have been obtained earlier.
A good outcome can include luck; a disappointing outcome does not automatically prove that the original reasoning was careless. Look for a specific adjustment to the next decision, such as testing an assumption sooner or involving someone with missing operational knowledge.
Keep confidential staff or customer details out of an unnecessarily broad shared journal. The useful record is the reasoning needed for learning, not a permanent collection of every personal detail discussed along the way.

