Welcome to Coach Kaizen
Most plants have the data. What's missing is the translation — turning raw exports into clear priorities your team can act on.
Upload your data, map your columns, and Coach Kaizen will walk you through a data quality review and Pareto analysis — step by step, with no surprises.
What would you like to analyze?
Upload your data file
Export your data as an Excel or CSV file, then upload it here. Coach Kaizen will detect your columns automatically.
Accepted formats: Excel (.xlsx, .xls) or CSV (.csv).
Click to choose a file, or drag and drop here
Excel (.xlsx) or CSV (.csv) accepted
Match your columns
Coach Kaizen made its best guess at matching your columns to standard fields. Review and adjust anything that looks off. Fields marked * are required.
What your Pareto will show
These two fields are highlighted in blue below — make sure they're mapped correctly before continuing.
Reason Code ★ = the column containing defect names or codes (e.g. "Surface Scratch", "Offset Error", "F07").
Scrap Quantity ★ = the column containing the number of units scrapped. If each row is one scrap event, map this to your count column.
Columns detected in your file
Map to standard fields
Data Quality Review
Before running your analysis, Coach Kaizen scans your data and scores it for completeness. This tells you how much to trust your results — before you act on them.
Overall data confidence
Field-by-field breakdown
Required fields directly affect your Pareto. Optional fields add depth but won't block your analysis.
Download your flagged data file
Get your original data back with problem rows highlighted — red for blank required fields, yellow for blank optional fields, orange for non-numeric quantities.
Coach note: These gaps likely exist in your system too — not just this export. Fixing them at the source means every future analysis gets cleaner automatically.
Truth Gate
The Truth Gate scores your data before analysis runs. It tells you how many rows are fully usable, what's dragging the score down, and what that means for your results. Your data is never changed here.
Confidence score breakdown
Coach's take
Before you continue
Your data confidence is below 70%. Results may be incomplete or point you toward the wrong priorities. We strongly recommend downloading your flagged file, fixing the gaps at the source, and re-uploading before acting on this analysis.
Results
Coach's Take
Data quality summary
See full quality details
Pareto chart — Top loss drivers
Top problems ranked
Click "Start Project" on any row to launch a DMAIC improvement project for that issue.
DMAIC Projects
Track your improvement projects through every phase of DMAIC.
No projects yet.
Run a Pareto analysis and click "Start Project" on your top problem, or start a project manually above.
Define
Clearly define the problem, the goal, and who's working on it.
Problem being addressed
Problem statement
Describe what is happening, where, how much, and since when — without assigning blame or suggesting a cause.
Example: "Surface scratches on Line 1 account for 18% of total scrap over the last 90 days."
Describe what is happening, where, how much, and since when. Do not assign blame or suggest causes yet.
Project goal
Your goal should be specific and measurable — include a target number and a deadline.
Example: "Reduce surface scratch scrap by 50% within 90 days."
Avoid vague goals like "reduce scrap" or "improve quality."
Project scope
Scope prevents "scope creep" — where a project grows beyond what's manageable. Be explicit about what's in and what's out.
Example: "In scope — Line 1 Day and Afternoon shifts. Out of scope — Night shift and Lines 2 and 3."
Team members
At minimum: someone who runs the process (operator), someone who owns the data (supervisor or quality), and someone from maintenance if equipment is involved.
Supervisors alone rarely find root causes. The people closest to the work know what's really happening.
Measure
Confirm your baseline and define how you'll track progress.
Baseline — current state
Your baseline is the "before" number. Without it, you can't prove your improvement worked. This number comes from your Pareto analysis — confirm it's accurate before moving forward.
Pulled from your Pareto analysis. Confirm or adjust if you have more current data.
Data collection plan
It answers: Who records the data, how it's measured, when it's recorded, and where it goes.
Also ask: is everyone measuring the same way? If operators define "defect" differently across shifts, your baseline is already skewed. This is called a Measurement System Analysis (MSA).
Analyze
Find the root cause — not just the symptom.
Your problem statement
Use this as your starting point for the fishbone and 5 Whys.
Fishbone session
Do this on a whiteboard with your team. Write your problem on the right. Draw 6 branches: Man, Machine, Method, Material, Environment, Measurement.
For each branch ask "what here could cause our problem?" Circle the top 2-3 causes — these feed your 5 Whys below.
Invite: operators, supervisors, maintenance, quality. The people closest to the work know what's really happening.
After your session, upload a photo of the completed diagram. Optional but recommended.
Upload a photo of your fishbone diagram
JPG, PNG, or PDF accepted
5 Whys
Start with your problem and ask "why did this happen?" Write the answer. Then ask "why did that happen?" Repeat until you reach a root cause you can actually fix.
The "therefore" check: Read your chain backwards using "therefore." If it makes logical sense, your chain is solid.
Watch out for: stopping at "operator error" — that's almost never the real root cause.
Start with your problem statement above and keep asking why. Use the "therefore" check to validate your chain.
Root cause conclusion
Based on your fishbone and 5 Whys, what is the confirmed root cause? It must be specific, actionable, and something your team can actually change. If you can't change it, it's not the root cause.
Improve
Design, test, and implement solutions that address the root cause.
Proposed solutions
Level 0 — Communication controls. Training, awareness, verbal instructions. Easiest to implement, least effective.
Level 1 — Output controls. Inspection and checking after the process. Catches defects but doesn't prevent them.
Level 2 — Input controls. Controlling inputs before they enter the process. More effective than output controls.
Level 3 — Error proofing (Poka-Yoke). The process physically cannot produce the defect. Hardest to achieve, most effective.
Ask for each solution: can we get this to Level 3?
List each solution you're considering or implementing. Note the level of control where possible.
Pilot plan
Running your solution on a small scale first protects your credibility. One shift, one line, one week. Document what happened, measure the result, and compare to your baseline before scaling.
How long to track: At least 2-3x your baseline measurement period. If baseline was 30 days, track for 60-90 days before declaring success.
Pilot result
Control
Lock in the gains and make sure the problem doesn't come back.
Control plan
· What the new standard is
· Who owns it
· How compliance is verified
· How often it's checked
· What happens when someone deviates
Audit cadence: Weekly for the first 30 days, monthly after. Without audits, standards drift — people get busy and old habits return.
Document the new standard. What changed, what must be maintained, and how?
Verify the improvement held
Upload a new data file from after your improvement was implemented. Coach Kaizen will map the columns, run the same Pareto, and show you a before/after comparison.
Upload post-improvement data file
Same format as your original upload — Excel (.xlsx) or CSV (.csv)
Confirm column mapping
Make sure these match how your original data was mapped so the comparison is accurate.
Before vs After — Pareto Comparison
Before Improvement
After Improvement
CI Templates
Interactive tools for your analysis. Fill them in-app, then print or download.