The four questions
What did we expect to happen. What actually happened. Why was there a difference. What will we do the same or differently next time.
That is the whole method. The discipline is in asking all four every time, in that order, and writing the answers down.
Start with the expectation, not the result
Most reviews start with what went wrong. Starting with the expectation changes the conversation. If the expectation was never written down, that is the first lesson. Decisions that are scored with an expected result before work starts are much easier to review.
Keep it about the work
A review is not a performance talk. The question is why the result differed from the expectation, and the answers are usually about information, timing or process, not effort. When people see that, they stop defending and start explaining.
When to run one
Run one when a decision reaches its due date, when a job goes noticeably better or worse than planned, and at the end of a season. Small, frequent reviews are more useful than one long review a year.
The following is a generic illustration, not a client. A plumbing shop expected a new after hours rule to cut missed calls. On the check date the owner and dispatcher compare the call log with the expectation, note that weekend calls still went unanswered, and set one new decision with an owner and a date.
Keep the lesson where the next decision will see it
A lesson that lives in someone's memory is lost when they are busy. North Star Vision keeps the gap between expected and actual on the decision record, so the next decision of the same shape starts from what happened last time.