Decisions
1 h

After-action review

Compare what was meant to happen with what happened, in a session structured so that the comparison can be honest.

Time cost
1 h
Output
A timeline, causes, and two or three owned changes.
Steps
6

Use when

  • A project, event or decision has concluded and the same kind will recur.
  • A team repeats mistakes it has already made.
  • Something went well and nobody knows why, which is as much a problem as failure.

Do not use when

  • Blame is on the table. The review will produce a defensible account rather than a true one.
  • Nothing similar will recur, so the lessons have nowhere to go.

Inputs required

  • The original intent, written down
  • What actually happened
  • The people involved, soon after

Procedure

  1. 01

    Run it soon

    Within days. Memory degrades toward a coherent story quickly, and the story is smoother than the events were.

  2. 02

    State the intent first

    What was supposed to happen, from the record rather than from memory. Doing this before discussing the outcome prevents the intent being rewritten to match it.

  3. 03

    Establish what actually happened

    Facts and sequence only. Settle the timeline before anyone explains it — most disagreements about why dissolve once the when is agreed.

  4. 04

    Ask why the difference

    Now interpret. Multiple causes, not one. Where a cause is a person, restate it as the condition that made the error likely.

  5. 05

    Ask what to sustain, not only what to fix

    What worked and why. Reviews that only look at failures systematically remove the reasons for success along with the problems.

  6. 06

    Write down two or three changes with owners

    Not fifteen. A review producing a long list produces no change, because nothing on it is anyone’s job.

Characteristic failure mode

It becomes a narrative. The group converges on a single clean story with one cause, usually a person, and the review’s output is agreement rather than learning.

Worked example

A community event drew half the expected attendance.

  1. 01Intent from the record: 200 attendees, promoted over six weeks.
  2. 02Actual sequence established: promotion started at week four, not six; the venue changed at week two.
  3. 03Causes: a late start and an unannounced venue change, not the weather everyone had blamed.
  4. 04Sustained: the volunteer rota, which worked and nobody had mentioned.

Result

Two changes with owners, and one practice explicitly preserved. The weather explanation, which the group had arrived believing, did not survive the timeline.

Where to go next