Leadership Katas
Focused, repeatable leadership exercises. Complete each kata; repeat until the move is rehearsed, not invented.
Kata 1: Executive Summary for an Outage
Time limit: 20 minutes. Goal: Write an exec-ready summary of a production incident, using BLUF (Concept 11). Setup: Use a real incident from the last 90 days or invent one at plausible scale.
Write a 5-sentence executive summary containing:
- What happened. One sentence.
- Customer impact. One sentence (scope + duration).
- Root cause or current understanding. One sentence.
- Mitigation and status. One sentence.
- Follow-up ask. One sentence ending in a specific decision or commitment request.
Constraints:
- No jargon that a non-technical VP would need to decode.
- No hedging ("it seems," "we believe we might") outside the root-cause sentence.
- No blame of individuals; systems and decisions only.
Repeat until: you can write this in under 10 minutes from a fresh prompt, and a non-engineering reader can correctly state (a) what happened, (b) the ask, and (c) the priority tier.
Kata 2: SBI Feedback on a Messy Code-Review Thread
Time limit: 15 minutes. Goal: Convert a vague or emotional piece of feedback into an SBI (Concept 14) delivery. Setup: Use a real code-review thread (with the names redacted) or the following template:
"The review went badly. A senior engineer was dismissive of the junior author, kept using phrases like 'obviously this won't work' and 'any engineer should know,' the author stopped responding, and the review stalled. I need to give feedback to the senior engineer."
Write the SBI-I feedback:
- Situation. One sentence with specific time and place.
- Behavior. Observable verbs. No adjectives about the person. 1-2 sentences with concrete examples.
- Impact. Named effect on author, team, timeline, or decision. 1-2 sentences.
- Intent ask. One question.
Then write the delivery - the 3-4 sentences you would actually say to the senior engineer, in a 1:1, not over Slack.
Red-flag checklist:
- No identity language ("you are dismissive," "you're not a team player").
- No mind-reading ("you were clearly trying to...").
- No sandwich (praise-critique-praise); trust the structure.
- Impact must include something observable, not only your feeling.
Repeat until: you can produce this in under 10 minutes for any scenario, and the delivery is something you would actually say out loud.
Kata 3: Roadmap Tradeoff Memo
Time limit: 25 minutes. Goal: Write a short memo defending a roadmap choice under capacity pressure. Setup: You are the Staff engineer on a platform team. Your manager has just been asked to add a major project to next quarter's Now. You have decided that adding it would force displacing the paved-road v1 initiative. Write the memo to your manager making the case.
The memo must contain:
- The capacity math. Show it. "Team of 4, 13-week quarter, 60% project time = ~31 engineer-weeks. Current Now = 28 eng-weeks. New project = ~15 eng-weeks. We are over by 12."
- The tradeoff named. "We can add the new project only by moving paved-road v1 to Next. Paved-road v1 slipping by a quarter means [named consequence]."
- A decision request. One of: (a) decline the new project, (b) accept and displace paved-road v1, (c) accept partially with a named scope reduction.
- Recommendation. Your position, with one sentence of reasoning.
Length: 300 words max. If it is longer, the thinking is not finished.
Repeat until: you can write this in under 15 minutes for any capacity-pressure scenario, and your recommendation is defensible under the question "what if we just work harder?" (answer: we already counted sustainable capacity; heroics are not a plan).
Kata 4: 3-Minute Architectural Story
Time limit: 10 minutes preparation, 3 minutes delivery. Goal: Deliver the six-part narrative arc (Concept 12) aloud, on a whiteboard or in a video call. Setup: Pick a system you know. Any size.
Prepare 3-6 bullets, one per slot:
- Problem. One concrete sentence.
- Constraint. One concrete sentence.
- Options. 2-3 alternatives, stated neutrally.
- Choice. One principle.
- Consequences. Plus + minus.
- Open questions. 2-3.
Deliver aloud in 3 minutes. Record yourself. Play it back.
Self-review:
- Can a listener restate the problem in one sentence?
- Did I name a constraint that actually ruled something out, or a pseudo-constraint?
- Did I state the alternatives fairly?
- Did I name at least one negative consequence?
- Did I name at least one honest open question?
- Did I finish under 3 minutes?
Repeat until: you can tell the story without notes, in under 3 minutes, on any whiteboard, for any system you know well.
Meta-Kata: Track Your Reps
Keep a short log. For each kata:
| Date | Kata | Scenario | What went wrong | What improved |
|---|
After 10 reps of each kata, review the log. Patterns will emerge. Those patterns are your personal weakest slot.
Completion Standard
- Kata 1: 3 different incidents, each under 10 minutes, each understandable by a non-engineer.
- Kata 2: 3 different scenarios, each with explicit SBI structure and no identity language.
- Kata 3: 3 different capacity-pressure memos under 300 words each.
- Kata 4: 3 different systems, each delivered aloud in 3 minutes, each with a named negative consequence.
- Meta-kata log has ≥12 entries and a named personal weakest slot.