Quality Engineering · Beginner · 22 min read
Root Cause Analysis: 5 Whys vs Fishbone Diagram
Move beyond symptoms. Learn when to follow one causal chain, when to explore many branches, and how to test a suspected cause before calling it the root cause.
Why this matters at work: weak root cause analysis creates weak corrective actions. A good analysis links the observed effect to a controllable system condition and then verifies that link with evidence.
Learning objectives
- Distinguish a symptom, direct cause, contributing cause, and root cause.
- Select 5 Whys, a fishbone diagram, or a combined approach based on problem structure.
- Build a defensible 5 Whys chain without forcing exactly five levels.
- Organize possible causes with the 6M fishbone categories.
- Prioritize and validate candidate causes before assigning corrective action.
Key concepts
The two tools solve different parts of the same problem. The 5 Whys develops depth along a causal path. The fishbone diagram develops breadth across multiple causal families.
Follow the chain
Best when the event sequence is reasonably clear and one answer can logically feed the next question.
Explore the system
Best when several functions, conditions, or mechanisms may contribute and the team needs structured brainstorming.
Go broad, then deep
Use the fishbone to generate candidate branches, then apply 5 Whys to the strongest branches.
Terminology
| Term | Meaning | Example |
|---|---|---|
| Symptom | The visible or measured effect. | Surface scratches increased after changeover. |
| Direct cause | The condition immediately producing the effect. | The side guide contacted the strip. |
| Contributing cause | A condition that increases probability or severity but may not be sufficient alone. | Poor lighting delayed detection. |
| Root cause | An underlying, controllable system condition whose removal prevents recurrence. | The product setup standard omitted the required guide clearance. |
| Candidate cause | A plausible explanation that still requires evidence. | Guide position drift. |
| 6M categories | Common fishbone families: Manpower/People, Method, Machine, Material, Measurement, and Mother Nature/Environment. | Measurement includes gauges, sampling, and inspection methods. |
Interactive tool selector
Select the characteristics that describe your problem. The recommendation updates as the problem structure changes.
Build a 5 Whys chain
Edit the problem and each answer. The chain visual updates immediately. Mark how strong the evidence is at every step.
Live causal chain
A strong chain requires a logical link and evidence at every transition.
Build a live fishbone diagram
Choose a 6M category, add or remove candidate causes, and watch the diagram rebuild itself without crowding the labels.
Live 6M fishbone
Click any branch to inspect its causes. The diagram organizes possibilities; it does not rank or validate them.
Focused on People.
Validate the candidate causes
Score the leading candidates using evidence strength, recurrence match, and controllability. The highest score identifies where to investigate first—not a proven root cause.
| Candidate cause | Category | Evidence | Recurrence match | Controllability | Score |
|---|
How to run a defensible root cause analysis
- Define the effect precisely. State what happened, where, when, how often, and how large the impact was.
- Contain before analyzing. Protect the customer and stabilize the process without confusing containment with corrective action.
- Select the tool. Use 5 Whys for a narrow chain, fishbone for broad exploration, or combine them.
- Separate facts from assumptions. Label each causal link as observed, partially supported, or unverified.
- Test candidate causes. Compare occurrence versus non-occurrence, reproduce the effect, review records, or run a controlled trial.
- Choose corrective action at the system level. Prefer standardization, mistake-proofing, design changes, alarms, and control-plan changes over reminders alone.
- Verify recurrence prevention. Monitor the outcome after implementation and confirm the effect does not return.
Worked example: recurring scratches after changeover
This example shows why the combined method is often stronger than using either tool alone.
Problem statement: During the first 30 minutes after product changeover, surface scratches increased from 0.4% to 3.1%, concentrated near the operator-side edge.
Common mistakes and analytical cautions
Documentation tools
These methods do not require specialized software. The priority is traceability: every cause, evidence source, owner, and decision should be visible.
| Tool | Best use | Minimum fields | Caution |
|---|---|---|---|
| Whiteboard or sticky notes | Facilitated fishbone brainstorming | Effect, categories, candidate causes, owner | Photograph and transcribe the final version. |
| Excel | 5 Whys log and validation matrix | Why level, answer, evidence, source, action, status | Do not convert scores into false precision. |
| PowerPoint / Visio | Clean fishbone communication | Effect, branch labels, prioritized causes | Keep evidence separate from brainstorming graphics. |
| Minitab Workspace | Structured quality projects and cause mapping | Problem statement, causes, tasks, evidence | Software structure does not replace process knowledge. |
| A3 / 8D report | Linking RCA to containment, action, and verification | Problem, analysis, root cause, corrective action, verification | Do not close the report before recurrence checks are complete. |
Practice challenge
Scenario: A laboratory turnaround time target is missed twice per week. Delays occur in sample receiving, preparation, testing, review, and result release.
Your task: use the selector above. Decide whether to start with 5 Whys, a fishbone, or both. Then add at least one cause under People, Method, Machine, Measurement, and Environment. Finally, identify one cause that could be tested using timestamp data.
Knowledge check
Your score: 0 out of 5.
Summary
- Define the effect first: precise problem statements produce stronger cause analysis.
- Use 5 Whys for depth: follow a logical causal chain and do not force exactly five levels.
- Use fishbone for breadth: organize many possible causes across the system.
- Combine the tools when useful: generate branches with the fishbone, then deepen the strongest branches with 5 Whys.
- Validate before acting: evidence and testing separate a plausible cause from a verified root cause.