Improvement and coaching kata

Improvement and coaching kata

When the direction is clear but the route is uncertain, practice learning in small steps. Understand the challenge, observe the current process and describe a near-term target condition. Select one obstacle and state what you predict will happen in the next safe test. Compare the actual result with the prediction and choose the next step from what was learned. The coach helps the learner think clearly rather than prescribing a hidden preferred answer. A target condition describes how the process should operate; it is not merely a list of actions. Keep the learning cycle connected to customer and process needs.

Follow the method

1. Challenge
2. Current condition
3. Target condition
4. Next experiment
5. Learn

Use a missed prediction to choose the next experiment

Fictional service team: learner Jo is improving request intake with coach Eva. The challenge is fewer requests waiting without a responsible receiver. In a six-request observation, four had a named receiver at handoff and two did not. Jo’s near-term target is that each request in the next agreed trial has a named receiver before leaving intake.

Follow the method

1. Current
2. Trial 1: receiver field
3. Observe blanks
4. Trial 2 planned

The reminder did not create an owner

Define a near-term operating condition, test one obstacle with a prediction and use a failed prediction to choose the next experiment.

Fictional service team: learner Jo is improving request intake with coach Eva. The challenge is fewer requests waiting without a responsible receiver. In a six-request observation, four had a named receiver at handoff and two did not. Jo’s near-term target is that each request in the next agreed trial has a named receiver before leaving intake.

Role: Improvement learner owning the experiment and coach supporting the reasoning.

Normal condition

The target describes the desired operating pattern and the next review horizon, not merely a distant performance number.

The gap

The receiving role is not always explicit at handoff.

• Use a bounded fictional trial, not an unapproved change to a live service.
• One small trial provides learning under its conditions; it does not establish permanent success.

1. Describe current work

   Jo records when receiver identity becomes known and which two handoffs lack it, rather than describing the team as poor at communication.

   Why: A current condition grounded in work gives the next test something specific to change.

   Evidence: Six-request record shows identity at the actual handoff.

2. Select one obstacle

   Jo focuses on visibility of receiving responsibility and predicts that a receiver field will produce six named handoffs. Eva asks what result would challenge that belief.

   Why: A prediction exposes the learner’s model before the outcome is known.

   Evidence: Trial 1 scope, change and expected six-of-six result are written first.

3. Compare expected and actual

   The fictional result is four named and two blank. Jo examines the two cases and finds that the field existed but role ownership was unresolved.

   Why: The failed prediction distinguishes a display problem from a responsibility problem.

   Evidence: Record the discrepancy without retrospectively lowering the prediction.

4. Choose the next test from learning

   Jo proposes a second bounded test in which a receiving counterpart agrees responsibility before handoff. The coach asks how this addresses the newly understood obstacle and when they can observe it.

   Why: The next experiment follows evidence, not a demand to make the first idea look successful.

   Evidence: Trial 2 plan names counterpart, cases, prediction and review; its future result remains blank.

A test improves the count by hiding cases

A proposed second trial excludes all requests with unclear receiving roles and achieves six named receivers among six included requests.

Report the exclusion and its limitation. Do not claim the original obstacle resolved; design a safe test that actually addresses the unclear-role cases.

Changing the population can make the count improve without changing the process that caused the gap.

Retain inclusion rules and excluded-case reasons beside the result.

Run a practice session

Materials

• Prediction/result cards
• Blank experiment record
• Learner/coach roles

1. Set operating target · 5 minutes

   What should happen at each handoff?

2. Reveal failed prediction · 7 minutes

   What changed in your understanding?

3. Practice due-date case · 9 minutes

   How does your next test address the actual obstacle?

4. Debrief scope · 5 minutes

   What does this small test not establish?

Debrief

• Ask for the learner’s prediction before showing results.
• Challenge a trial that excludes the very cases it should teach about.

Cover the actual column, write a prediction, then compare and choose a next test before reading the answer.

Transfer into the work

Owner: Improvement learner with an identified coach and process owner

Record: Current/target condition and experiment learning record

Review: At the agreed next observation after each bounded test

Evidence: Prediction, actual, inclusion conditions and resulting knowledge

Revisit the obstacle or observation method; keep the target and local authority visible while adapting the next test.