OperationsFramework / modelModelAccessible

Root cause analysis/Pareto analysis

How can root cause analysis/pareto analysis improve people, teams, or organisational effectiveness?

AccessibleOperationalIndividual2 min read

Contents

Root cause analysis (RCA) is a class of problem-solving methods aimed at identifying the root causes of problems or events.

Root cause analysis (RCA) is a family of methods for explaining why an unwanted event or pattern occurred and identifying changes that can prevent recurrence. Pareto analysis helps prioritise categories by frequency, cost or consequence. A fishbone diagram helps organise possible causes. These tools complement one another, but none alone proves a root cause.

When to use it

Use RCA when a recurring defect, failure, delay, incident or outcome has material consequences and treating the symptom has not been sufficient. Scale the investigation to severity and learning potential.

Use Pareto analysis when many categories compete for attention and a consistent measure can show where the largest share of effect occurs. Use a fishbone diagram early to widen causal inquiry across people, process, equipment, materials, environment, measurement and management-system factors.

(a) Ishikawa diagram, or cause and effect diagram. (b) Pareto diagram

Origins

RCA does not descend from one diagram. It draws on engineering failure analysis, quality management, safety investigation and systems thinking. Kaoru Ishikawa popularised the cause-and-effect or fishbone diagram in Japanese quality practice; it became widely used in the 1960s and is counted among the seven basic quality tools. Pareto charts derive from Vilfredo Pareto’s observations and Joseph Juran’s later application of the “vital few” idea to quality.

What it is

An RCA moves from a clearly defined event to evidence about causal mechanisms and control weaknesses. A root cause is not simply the last error in a chain or the person nearest the failure. It is a condition that, if changed, would materially reduce recurrence or consequence.

A fishbone diagram structures hypotheses; a Pareto chart ranks observed categories; process maps, timelines, change analysis, barrier analysis, interviews, tests and data analysis evaluate what actually happened. The output should distinguish confirmed causes, contributing conditions and unresolved hypotheses.

How to use it

Form a team with process knowledge, investigation skill and enough independence to challenge assumptions. Include affected workers, customers, suppliers or specialists where relevant, and protect psychological safety.

Define the event, expected state, actual state, scope and consequence. Preserve evidence before changing the process. Build a timeline and map the work as performed, not only as documented.

Generate causal hypotheses with a fishbone diagram, repeated “why” questions or other structured methods. Look beyond individual action to design, workload, incentives, training, interfaces, maintenance, information, supervision and barriers.

Test each important hypothesis against records, observation, experiment or triangulated testimony. Ask whether the condition preceded the event, whether a mechanism connects it to the outcome and whether alternative explanations fit the evidence.

Use a Pareto chart to prioritise recurring categories by an appropriate consequence measure. The familiar claim that 80 per cent of effects come from 20 per cent of causes is a heuristic, not a law; calculate the actual distribution.

Develop corrective actions that remove or control confirmed causes. Prefer stronger design, engineering and process controls over reminders alone. Assign owners and dates, assess new risks and verify effectiveness after implementation.

Final analysis

RCA is not a single standard method, and different domains require different investigative depth. Begin with high-consequence or recurring problems, but preserve mechanisms for learning from near misses and weak signals.

The greatest value comes from closing the loop. A plausible diagram and action list are not proof. Verify that the intervention was implemented, changed the causal condition and reduced recurrence without shifting harm elsewhere.

Top practical tip

Write the problem as a precise gap between expected and observed performance, with time, place, scope and impact. A well-bounded problem makes evidence collection and causal testing far more reliable.

Top pitfall

Do not stop at “human error,” vote on causes in a workshop or treat the largest Pareto bar as causation. Investigate why the system allowed the action, test mechanisms and verify the corrective control.