When should a plant perform a formal root cause analysis?
A formal root cause analysis is justified when consequences, recurrence or learning value warrant a structured investigation. Define triggers before incidents occur so significant failures are treated consistently.
Set objective investigation triggers
Investigating every minor defect formally can overwhelm a plant, while investigating only catastrophic events misses preventable patterns. Define thresholds that consider credible potential consequences as well as actual loss.
- Safety and environmental significance: Actual or credible serious harm, loss of containment or failure of a protective barrier.
- Major business disruption: Events exceeding a defined downtime, cost, quality or production threshold.
- Repeat failure: The same failure mechanism returns despite previous corrective work.
- Critical asset deterioration: A condition finding reveals a serious systemic weakness before failure.
- Recurring process breakdown: Similar procedural, material or maintenance control failures occur across assets.
Site thresholds should be approved by leadership and relevant technical and safety authorities. Regulatory and mandatory investigation requirements take precedence.
Choose the right investigation depth
A minor isolated event may need a documented field review. A repeat defect may need a facilitated cause tree and evidence review. A major event may require multidisciplinary participation, preserved evidence, independent oversight and formal action tracking. Escalate if initial findings reveal more severe risk than expected.
Preserve evidence early
Secure failed components when safe and permitted. Collect photographs, alarm histories, operating conditions, work orders, drawings and witness accounts. Build a timeline of verified observations. Keep hypotheses separate from facts. Starting with “operator error” or “bad part” can hide design, procedure, training and management system weaknesses.
Look for more than one cause
Determine the physical failure mechanism, contributing decisions and organizational controls that allowed the condition. Multiple causes can interact. Five Whys and fishbone diagrams help structure thinking but do not replace evidence. A single causal chain is rarely enough for a complex incident.
Example of a repeat failure
A process pump experiences three seal failures in one year. Replacing the seal restores function but does not explain recurrence. The team examines operating data, seal condition, flush arrangement, installation and procedures. Possible causes are tested rather than assumed. This is an illustrative scenario, not a diagnosis.
Close and verify corrective actions
- Define actions that address evidence backed causes, not vague reminders.
- Assign an owner, deadline, acceptance criteria and required proof.
- Update design, procedures, training or maintenance strategies when justified.
- Verify both implementation and effectiveness after an appropriate period.
- Share relevant lessons with other assets and sites.
Measure repeat events, overdue actions and verified effectiveness. The number of RCA reports alone is not evidence that reliability improved.
Explore the assessment
KANE can help evaluate the underlying maintenance process and build a practical improvement plan.
View the related assessment →Related answer: How can a plant identify repeat equipment failures?