# Retrospectives and Kaizen

Turn reflection into one or two owned improvement experiments that are reviewed next time.

Duration: 10 min

A retrospective applies empiricism to the team itself. Psychological safety makes honest evidence possible; a small owned experiment closes the loop.

## fmap-m4-l4-v1

The approved narration and storyboard are present in the source pack. No unreviewed provider media is activated.

## fmap-m4-l4-digest

### The improvement engine

Kaizen is continuous improvement through small, repeated changes. A retrospective puts that idea on the calendar: inspect the way of working, decide what to try, and review the result next time.

Three conditions keep the practice useful. Psychological safety makes it possible to report inconvenient evidence. The prime-directive spirit directs attention toward conditions rather than easy blame. Most importantly, the meeting ends with one or two concrete experiments, each owned by someone and reviewed at the next retrospective.

Formats are a menu rather than a doctrine:

- **Start–Stop–Continue:** decide what to begin, stop, and preserve.
- **4Ls:** explore what the team liked, learned, lacked, and longed for.
- **Sailboat:** identify winds, anchors, rocks, and the destination.
- **Mad–Sad–Glad:** surface emotions when they are the evidence the group is avoiding.

Rotate formats before answers become automatic. The format opens the conversation; the experiment creates improvement.

A complaint session with no changed behavior is not a successful retrospective. Venting can be useful, but it becomes part of an improvement loop only when the team chooses a safe experiment, names an owner and review date, then keeps, adjusts, or drops it based on evidence.

## fmap-m4-l4-r1

### Four retrospective formats

Pick one and rotate often. No format is the retrospective; each is a different way to gather evidence for an improvement decision.

| Format                  | Questions                                        | Best when                                                |
| ----------------------- | ------------------------------------------------ | -------------------------------------------------------- |
| **Start–Stop–Continue** | What should we start, stop, and keep doing?      | Most situations; fast and decision-oriented              |
| **4Ls**                 | What was liked, learned, lacked, and longed for? | A demanding period needs reflection and learning         |
| **Sailboat**            | What was wind, anchor, rock, and destination?    | The team needs to surface risks as well as help and drag |
| **Mad–Sad–Glad**        | What produced anger, loss, or satisfaction?      | Emotion is important evidence that has remained unspoken |

#### A 30-minute agenda

1. **5 minutes—set the stage.** Revisit the last experiment and check whether people can speak safely.
2. **10 minutes—gather evidence.** Use one format and ask for observations rather than accusations.
3. **10 minutes—decide.** Select one or two small experiments rather than an unowned wish list.
4. **5 minutes—close the loop.** Name each owner and the date when the team will review the result.

Assume people acted as reasonably as they could with the information and conditions available. Investigate the system before blaming an individual. The defining output is an experiment somebody owns.

## fmap-m4-l4-i1

Your response remains local to your browser. It is ungraded, has no minimum length, and may be skipped without penalty.