What is root cause analysis?
Root cause analysis (RCA) is a structured investigation into why a problem happened, traced past the immediate trigger to the condition that allowed it. A conveyor jam’s immediate cause may be a crushed carton, and its root cause a supplier’s switch to thinner cardboard. Clearing the jam restores the line, while changing the carton specification prevents the next one.
Common techniques include the five whys, fishbone diagrams that sort possible causes into categories, fault tree analysis, and Pareto charts that rank causes by frequency. Hospitals use RCA after serious patient-safety events and IT teams after outages. AI speeds up the evidence gathering. Anomaly detection and diagnostic analytics show which readings or settings changed when the problem began, and language models group similar incident reports.
Root cause analysis in practice
A managed service provider saw a rise in dropped video calls across several clients. An analyst had an assistant group the tickets by symptom, site, and start time and line them up against the change log. The cluster matched a firmware update on one model of wireless access point. Rolling it back closed the tickets, and the provider now rolls out firmware in stages.
What to watch for
- A model shows what changed alongside the failure, and a person must confirm the cause, ideally by testing the fix.
- An investigation that ends at “human error” has usually stopped too early, since the useful question is what made the error easy to make.
- Track whether corrective actions were completed and whether the problem came back.