Analyze phase: root cause analysis and waste identification
Analyze is where the team stops listing symptoms and proves causes. The exam tests your use of 5 Whys, fishbone diagrams, Pareto and stratification, your ability to verify a cause with data, and your ability to see the eight wastes inside a process map.
Root cause toolkit
- 5 Whys: repeatedly ask why from symptom to systemic cause; stop where action is possible and evidence supports the link.
- Fishbone (Ishikawa): branches of the 6M - Man, Machine, Material, Method, Measurement, Mother Nature/environment - generate candidate causes; they do not prove them.
- Pareto: ranks causes by frequency or cost, reflecting the 80/20 idea; show counts as well as percentages.
- Scatter plot: shows association between a suspected x and the Y - correlation, not proof of causation.
- Stratification: splits data by shift, machine, supplier or operator to expose hidden differences.
- FMEA: scores Severity × Occurrence × Detection to prioritise which cause deserves attention first.
Verify before you fix
A candidate becomes a root cause only when evidence links it to the problem: a stratified Pareto, a controlled comparison, or a change that shifts the defect rate. Distinguish the occurrence cause (why the defect was made) from the escape cause (why it reached the customer); both may need action.
The eight wastes
Overproduction, Inventory, Waiting, Motion, Transport, Overprocessing, Defects and unused Skills (TIMWOODS). Value test: the customer would pay for it, it is done right first time, and it changes form, fit, function or information.
Worked example
A warehouse reports 120 late shipments a month. Pareto shows 78% come from one carrier lane; stratification shows 90% of those occur on the 06:00 pick. 5 Whys: late van → pick started late → pick list printed at 05:45 → printer shared with returns → returns queue peaks before 06:00. Timestamps confirm the queue, and no other lane shows the pattern. Root cause: shared printer queue, not "careless pickers."
Traps that cost marks
- Stopping at "operator error" - a cause that explains everything explains nothing.
- Picking the loudest brainstormed cause without data.
- Reading a scatter plot as proof of causation.
- Calling necessary inventory or transport value-adding.
- Confusing a defect with its escape point.