Please enable JavaScript.
Coggle requires JavaScript to display documents.
Chapter 14 – Analyze Phase (DMAIC) - Coggle Diagram
Chapter 14 – Analyze Phase (DMAIC)
Purpose of the Analyze Phase
Convert measured data into meaningful insights
Identify true root causes of defects or variation
Validate causes before jumping to solutions
Build confidence that improvements will work
Role of the Team
Black Belts and Green Belts lead the analysis work
Team members and subject matter experts provide insight
Findings are reviewed and validated collaboratively
Analysis supports decisions for Improve phase
Root Cause Analysis
Used when causes are not immediately obvious
Focuses on identifying primary drivers of problems
Can build on work from Define and Measure phases
Requires verification before acceptance
Cause-and-Effect (Fishbone) Diagram
Visual tool for brainstorming possible causes
Encourages structured thinking and team participation
Organizes ideas into logical categories
Common Fishbone Categories
People involved in performing the process
Process or machines transforming inputs to outputs
Procedures or methods used to complete work
Materials used as inputs
Equipment supporting the process
Environment surrounding the process
How Fishbone Is Used
Start with a clearly defined problem or defect
Brainstorm causes under each category
Use sticky notes or free movement of ideas
Apply the 5 Whys to reach deeper causes
Remove weak or unsupported causes
Highlight most likely root causes
Root Cause Verification
Confirms whether suspected causes are real
Prevents teams from fixing the wrong problem
Uses data, observation, or analysis methods
Root Cause Verification Matrix
Lists problem and possible root causes
Documents verification method used
Explains why the method was chosen
Records results and verification outcome
Creates accountability and transparency
Graphical Analysis Tools
Used to explore relationships in data
Help visualize variation and patterns
Often serve as a starting point for deeper analysis
Pareto Charts
Based on the 80/20 principle
Identify the few causes creating most of the impact
Help teams focus on highest-value improvement areas
Can be drilled down for deeper analysis
Box Plots (Box-and-Whisker Charts)
Compare distributions between groups
Show median, variation, and outliers visually
Useful for comparing operators, shifts, inputs, or methods
Help identify meaningful differences without complex math
Key Box Plot Elements
Middle line shows the median value
Box represents middle 50 percent of data
Whiskers show typical data range
Dots represent statistical outliers
Hypothesis Testing
Tests assumptions about population behavior
Uses sample data to draw conclusions
Includes risk of incorrect conclusions
Results expressed using p-values
Statistical Analysis Overview
Supports conclusions drawn from samples
Quantifies confidence and risk of error
Usually performed using statistical software
Correlation and Regression Analysis
Examines relationships between variables
Determines whether inputs influence outputs
Regression quantifies strength of relationships
Requires numerical (continuous) data
Design of Experiments (DoE)
Used when multiple variables interact
Tests factors systematically and efficiently
Provides high confidence before implementation
Often used to finalize Improve-phase decisions
Analyze Phase Tollgate Checklist
Root causes have been clearly identified
Causes are prioritized based on impact
Statistical evidence backs key assumptions
Relationships between variables are understood
Sponsor agrees team is ready to Improve