KPIs, strategy & continuous improvement
Root cause analysis tools: 5 Whys, fishbone, Pareto and more
How to use 5 Whys, fishbone diagrams, Pareto charts, process maps and data checks to find root causes, with business examples and common mistakes.
When a KPI is red, the natural reaction is to fix the symptom: recount the stock, redo the payroll, apologize to the customer. Root cause analysis (RCA) asks why the problem happened, so the fix prevents it from happening again. The tools are simple; the discipline is in using them with data and the people who do the work.
The core tools
| Tool | What it does | Best for |
|---|---|---|
| Pareto chart | Ranks causes by frequency or impact | Deciding which problem to tackle first |
| Fishbone (Ishikawa) diagram | Organizes possible causes by category | Brainstorming causes as a team |
| 5 Whys | Asks "why" repeatedly to move from symptom to cause | Simple to moderate problems |
| Process map | Shows steps, handoffs and decisions | Problems that occur at handoffs |
| Data check (stratification) | Splits data by branch, shift, product, person | Testing which causes are real |
Pareto: pick the vital few
List problem occurrences by cause and sort from most to least frequent. Usually a few causes account for most occurrences.
Example: 400 late deliveries in a quarter.
| Cause | Count | Cumulative % |
|---|---|---|
| Stock not available at picking | 164 | 41% |
| Truck capacity shortfall | 108 | 68% |
| Late release from credit hold | 72 | 86% |
| Wrong address or contact | 32 | 94% |
| Others | 24 | 100% |
The first two causes explain 68% of late deliveries. Start there.
Fishbone: structure the brainstorm
Draw the problem as the "head" and causes as "bones" in categories. A common set for operations is people, methods, machines, materials, measurement and environment; for services, policies, procedures, people, systems, place and suppliers. The team lists possible causes under each, then decides which to test with data.
5 Whys: go deeper
Problem: payroll corrections after release average 45 per cutoff.
- Why? Overtime was paid incorrectly.
- Why? Overtime approvals arrived after cutoff.
- Why? Supervisors approve in batches at the end of the period.
- Why? They get no reminder and approvals happen in chat.
- Why? There is no approval workflow with deadlines.
Root cause: no structured approval process with deadlines. Countermeasure: approval workflow with reminders before cutoff (payroll cutoff best practices).
5 Whys works best when answers are verified with facts at each step, not assumptions.
Verify with data
A cause on a fishbone is a hypothesis. Test it: stratify the data by branch, shift, product, supplier or person, and see whether the problem concentrates where the suspected cause exists. Fixing an unverified cause wastes time and erodes trust in the method.
Running an RCA session
- Define the problem precisely, with numbers and scope.
- Bring the people who do the work, not only managers.
- Use Pareto to choose focus, fishbone to list causes, 5 Whys to go deeper.
- Agree which causes to verify and how.
- Verify, then decide countermeasures with owners and dates.
- Check later whether the KPI improved.
Common mistakes
- Stopping at "human error" (ask why the process allowed the error)
- Blaming individuals instead of fixing systems
- Skipping verification
- One-person analysis without the people closest to the work
- No follow-up to check the fix worked
Where root cause analysis fits
RCA sits inside larger methods: the Analyze phase of DMAIC, the Plan step of PDCA, the investigation in CAPA and the left side of an A3 report.
Where Cirquolus fits
Cirquolus KPI software connects red measures on scorecards to operational reviews and the improvement portfolio, where DMAIC, PDCA, CAPA, Kaizen and FMEA workspaces hold the analysis, countermeasures and follow-up.
Frequently asked questions
What is root cause analysis?
A structured approach to finding the underlying causes of a problem so fixes prevent recurrence, rather than only treating symptoms.
What are the most common root cause analysis tools?
Pareto charts, fishbone diagrams, 5 Whys, process maps and data stratification.
Is "human error" a root cause?
Rarely. Ask why the process allowed the error to happen or go undetected, and fix that.
When should root cause analysis be done?
When a problem recurs, has significant impact or its cause is unclear.
Bring this workflow to a Cirquolus demo.
Show us your current process, companies and branches. We will map it to the modules you need and propose a realistic rollout.