Apply Kaizen continuous improvement — run 5 Whys analysis, map cause-and-effect, and propose incremental fixes to any process or workflow. Works from retro transcripts, standup patterns, or described problems. Trigger on: kaizen, continuous improvement, root cause, 5 whys, process improvement, why does this keep happening.
Apply Kaizen continuous improvement — run 5 Whys analysis, map cause-and-effect, and propose incremental fixes to any process or workflow. Works from retro transcripts, standup patterns, or described problems. Trigger on: kaizen, continuous improvement, root cause, 5 whys, process improvement, why does this keep happening.
dependencies
[]
connectors
[{"name":"Slack","reason":"Pull team discussions about recurring problems and frustrations"},{"name":"Linear","reason":"Reference issue history and cycle data to identify bottleneck patterns"},{"name":"Jira","reason":"Pull sprint data, bug trends, and blocked tickets for process analysis"},{"name":"Notion","reason":"Reference existing process docs and past improvement initiatives"}]
author
Harmony
version
1.0.0
icon
sparkles
cover
screenshots
[]
category
Productivity
Purpose
The Apply Kaizen Method skill transforms recurring frustrations and bottlenecks into concrete, incremental improvements. Kaizen (continuous improvement) is rooted in the principle that small, structured changes accumulate into major gains. This skill runs a disciplined root-cause analysis using the 5 Whys method and Cause-and-Effect (fishbone) mapping, then proposes 3–5 low-risk, high-impact fixes scoped for a 2-week sprint.
Unlike broad process redesigns, Kaizen focuses on the actual conditions teams face today—data from recent retros, standup chatter, ticket patterns, or user feedback. The output is actionable: each improvement has a success metric, an owner, and a clear first step.
When to Use
Trigger Phrases:
"Why does [X] keep happening?"
"We need better [process]"
"I want continuous improvement around [topic]"
"What's the root cause of our bottleneck?"
"Let's do a Kaizen sprint on [workflow]"
"Why are we still stuck on this?"
Common Scenarios:
Operations/Engineering: Long customer onboarding cycles, repeated production incidents, deployment friction.
Product/Design: High volume of rework feedback, slow design-to-dev handoff, unclear requirements.
Sales/Support: Long ticket resolution times, missed follow-ups, inconsistent customer data entry.
Leadership/Teams: High meeting load, unclear decision-making, duplicate work across squads.
Problems where blame is the real issue (Kaizen assumes systems, not people, are the problem).
Instructions
Step 1: Define the Problem Statement
Clarify the single, measurable problem you're solving. Avoid multi-issue sprawl.
Questions to ask:
What is the current state (time, cost, defects, wait)?
What is the desired state?
How often does this occur?
Who feels the pain most acutely?
Output: A 1–2 sentence problem statement. Example: "Customer onboarding (trial signup → first automated report) averages 3 weeks; we target 5 business days."
Step 2: Gather Evidence
Pull data from Harmony sources: Slack channels, Linear issues, Jira sprints, Notion retros, or user feedback. Look for:
Frequency: How many times has this been reported/observed?
Duration: How long does the process take (vs. ideal)?
Cost: What is lost (time, money, morale, customer satisfaction)?
Patterns: Does it happen at specific points or with certain people/tools?
Action: Surface at least 3 concrete data points. "Retro notes from Jan–Feb show 7 mentions of slow design handoff. Linear shows 6 blocked product tasks this cycle, averaging 4-day wait for design review."
Step 3: Run 5 Whys Analysis
Ask "Why?" five times to drill past surface symptoms to root causes. Each layer should answer the previous "why."
Template:
Why does [problem] occur?
Why does [that reason] happen?
Why does [that] happen?
Why does [that] happen?
Why does [that] happen? (Usually here you hit process, system, or awareness gaps.)
Rule: Each answer must be provable or testable from your gathered evidence. Avoid speculation.
2-Week Sprint Plan (week-by-week roadmap, owners, checkpoints)
Next Kaizen Cycle (what to measure/learn for the next round)
Example: Veltrix Customer Onboarding (B2B SaaS)
Problem Statement:
Veltrix (AI data platform, $2M ARR) launched a trial program 8 weeks ago. Customer onboarding (trial signup → first automated report) averages 3 weeks; product goal is 5 business days. This gap is causing 40% of trials to go dark before activation.
Evidence & Data
Retro notes (Feb): 12 team mentions of "slow onboarding," "customers asking where their data is," "support queue flooded with where's my report questions."
Linear: 8 blocked customer support tasks this cycle (avg 4–6 day wait). Root issue tags: "waiting for CS config," "unclear customer data format," "manual report gen."
Slack #support: 23 messages in past 2 weeks asking "when will customer X get their first report?"
HubSpot: 41 new trial signups in Feb; only 11 activated; 18 dormant >7 days; 12 churn before week 1.
5 Whys Drill
Why do trials take 3 weeks to activate?
→ Because Veltrix CSMs manually build each customer's data pipeline and test the first report before enabling.
Why do CSMs do this manually?
→ Because the product doesn't have a self-service data upload wizard; customers get stuck when they don't know data format/schema.
Why doesn't the product have a wizard?
→ Because the API documentation is incomplete and the UI is designed for power users, not first-time trial users.
Why is onboarding designed for power users?
→ Because the product team built for existing enterprise customers; trial users were added later without re-designing the flow.
Why wasn't the flow re-designed for trials?
→ Because there was no feedback loop between support (handling trial issues) and product (building features). No one tracked trial activation as a metric until recently.
Cause-and-Effect Map
[3-week trial activation lag]
|
People | Tools
------ | -----
- CSM knowledge only ----+ | +---- No self-service wizard
- No escalation path ----+ | +---- Incomplete API docs
- Support overwhelmed ---+ | +---- UI assumes expert users
|
Process | Environment
------- | -----------
- No handoff from sales --+ | +---- No trial SLA
- Manual config --------+ | +---- No activation dashboard
- No feedback loop prod-support + +---- Support doesn't track activation %
Root Cause Scoring
Root Cause
Frequency
Impact
Score
Priority
No self-service wizard (tools)
5
5
25
CRITICAL
Manual config by CSM (process)
5
4
20
HIGH
No feedback loop (people)
4
4
16
HIGH
Incomplete API docs (tools)
4
3
12
MEDIUM
Support overwhelmed (people)
3
3
9
MEDIUM
Proposed Kaizen Improvements
Improvement 1: Launch Quick-Start Template
What: Create a 5-minute guided template inside the product UI that auto-generates a sample data pipeline for new trials.
Why it fixes: Removes manual CSM config for 70% of trial types; lets customers self-serve.
Owner: Sarah (Product Lead)
Timeline: Start Mon, ship Thu.
Success Metric: 60% of new trials use template; CSM config time drops from 2 hours to 30 min.
Improvement 2: Auto-Generate First Report
What: When customer data passes validation, trigger an automated report generation (pre-configured dashboard snapshot) instead of waiting for CSM to hand-craft.
Why it fixes: Cuts manual bottleneck; delivers value within 24h of data upload.
Success Metric: First report generated within 4 hours of data upload; support tickets drop 30%.
Improvement 3: Sync Support Tickets → Product Backlog
What: Set up a weekly 30-min sync between support lead (Marcus) and product lead (Sarah). Marcus surface top 3 trial friction points; Sarah threads into product backlog.
Why it fixes: Closes feedback loop; product sees real trial pain; prioritizes high-impact fixes.
Owner: Marcus (Support Lead)
Timeline: Start next Mon, recurring weekly.
Success Metric: Top 3 support friction points addressed in next cycle; support queue <1 day avg wait.
Improvement 4: Publish Simplified API Guide for Trials
What: Create a 2-page "trial data format" guide with 3 copy-paste examples (CSV, JSON, webhook). Post in help center and send with trial welcome email.
Why it fixes: Removes confusion about data schema; supports self-serve path.
Owner: Deena (Technical Writer)
Timeline: Start Wed, ship Fri.
Success Metric: 50% of new trials reference guide before contacting support.
Improvement 5: Add Trial Activation Dashboard
What: Create a Notion dashboard CSMs and product check daily: trial signups, data uploaded, reports generated, activation status (trial → paid, churned, dormant).
Why it fixes: Makes progress visible; surfaces stuck customers early; enables rapid intervention.
Owner: Sarah (Product) + Marcus (Support)
Timeline: Start Tue, ship Wed.
Success Metric: CSMs identify stuck trials within 24h; re-engage 20% of at-risk trials before churn.
Thu 2pm: Retro + metrics review. Did we hit targets? What's working? What needs another cycle?
Fri: Document learnings. Plan next Kaizen cycle (e.g., "Improve paid plan onboarding," "Reduce support volume on X topic").
Next Kaizen Cycle
After this sprint, measure:
Did trials using template activate 3x faster?
Is support queue back to <1 day?
Which trial customer profiles still churn early? (Next cycle focus.)
What's the next high-pain process to improve?
Recommended follow-up: In 2 weeks, run a second Kaizen sprint on "Reduce trial-to-paid friction" (now that activation is fixed, why don't 50% convert?).
Edge Cases
Edge Case 1: Feedback is vague ("We're just slow")
→ Push back gently. Ask: "Can you point me to one customer who complained?" or "Show me a ticket that was painful." Kaizen requires data, not feelings.
Edge Case 2: Root cause requires re-architecture
→ That's not Kaizen; that's a bigger project. Call it out: "This fix requires 6 weeks and new staffing. Let's find a smaller Kaizen win first, then propose the re-arch separately."
Edge Case 3: Team says "We've tried this before"
→ Ask what was different last time. Was it a different tool, wrong owner, no buy-in, or wrong metric? Kaizen often fails because implementation wasn't rigged to succeed, not because the idea was wrong.
Edge Case 4: Multiple bottlenecks, can't prioritize
→ Pick ONE problem statement. Run the full analysis on that. Ship it. Then do the next Kaizen sprint on the second problem. Spread it across 3–4 sprints if needed.
Edge Case 5: Success metric is hard to measure
→ Pick a proxy. Can't measure "morale"? Measure "avg support tickets per week" or "% team saying process is clearer in survey." Perfection is the enemy of progress.
