5 Root Cause Analysis Mistakes That Kill Your Investigations

After reviewing thousands of incident investigations across mining, construction, and manufacturing, a pattern emerges: the same mistakes show up again and again.

These aren't mistakes made by careless investigators. They're made by experienced safety professionals who genuinely want to prevent incidents. The mistakes are systemic — baked into how most organizations approach investigation.

Here are the five most common mistakes — and how to avoid them.

Mistake #1: Stopping at "Human Error"

This is the most common — and most damaging — investigation mistake. When your root cause is a variation of "the worker messed up," you haven't found a root cause at all. You've found a symptom.

Human error is always the proximate cause of incidents. Someone did something (or didn't do something) that led to the event. But that tells us nothing useful.

The question isn't what they did wrong. The question is why was it possible for them to do it wrong?

Why This Happens

• It's the easiest conclusion — you can see the error clearly
• It feels satisfying — someone to "blame"
• It's fast — investigation can close quickly
• It protects the organization from harder questions

Mistake #2: Writing Weak Corrective Actions

If retraining solved problems, we'd have zero incidents by now. Workers are trained constantly. And yet incidents keep happening.

Retraining, adding warning signs, and "reminding workers to be careful" are all administrative controls — the weakest level on the hierarchy of controls. They rely on perfect human behavior, every time, forever.

That's not how humans work.

The Hierarchy of Controls (Most → Least Effective)

1. Elimination: Remove the hazard entirely
2. Substitution: Replace with something less hazardous
3. Engineering controls: Physically prevent exposure
4. Administrative controls: Change how people work
5. PPE: Protect the individual

Most corrective actions cluster at levels 4 and 5. That's why incidents recur.

Mistake #3: Confirmation Bias

It's human nature. We form hypotheses quickly (often within minutes of hearing about an incident) and then seek evidence that confirms them. Evidence that contradicts our hypothesis? We minimize it, explain it away, or don't notice it at all.

The result: investigations that feel thorough but miss critical contributing factors.

Signs of Confirmation Bias

• You knew the "root cause" within the first hour
• All the evidence points the same direction (too neatly)
• Witnesses who disagreed with your theory were "mistaken"
• The conclusion matches what management expected

Mistake #4: Not Reaching Organizational Factors

Equipment doesn't maintain itself. Procedures don't write themselves. Training doesn't happen automatically. Behind every immediate cause is an organizational factor — a decision, system, or cultural element that allowed the conditions to exist.

These are the true root causes. And most investigations never reach them.

Why Investigations Stop Short

• Organizational factors are uncomfortable — they implicate management
• They're harder to identify than physical causes
• They require deeper analysis and more time
• There's often organizational pressure to keep it simple

Mistake #5: Investigating in Isolation

Organizations experience patterns of incidents — similar contributing factors appearing again and again. But because each investigation is siloed, the pattern remains invisible.

How many times has your organization had incidents where:

• Shift handover communication was a factor?
• Time pressure contributed?
• Contractor management was involved?
• The procedure was outdated or impractical?

If you can't answer that immediately, you're investigating in isolation.

The Meta-Mistake: Treating Investigation as Compliance

Underlying all five mistakes is a deeper issue: many organizations treat incident investigation as a compliance exercise rather than a learning opportunity.

When the goal is to complete the form, close the case, and satisfy the regulator, investigations become superficial. When the goal is to genuinely understand what happened and prevent recurrence, investigations become powerful.

"The quality of your investigations reveals whether your organization truly wants to learn — or just wants to look like it's learning."

How to Fix Your Investigation Process

1. Adopt ICAM methodology: It's specifically designed to reach organizational factors.
2. Train investigators properly: Investigation is a skill, not just common sense.
3. Allow adequate time: Rushing guarantees shallow results.
4. Create psychological safety: Investigators need freedom to report uncomfortable findings.
5. Track and trend: Connect the dots across multiple incidents.
6. Measure corrective action effectiveness: Did it actually work?

Building a Better Investigation Culture

Lasting improvement requires more than fixing individual investigation reports. It requires building an environment where honest incident reporting is rewarded, investigators have the time and training to go deep, organizational factors can be named without political consequences, and corrective actions are tracked until they demonstrably reduce risk. That's a culture question as much as a methodology question — and it starts with how senior leaders respond when an investigation implicates management systems rather than worker behaviour. When leaders protect the process, the process improves. When they protect their own decisions, investigations stay shallow. Investing in structured methodology — and the training, time, and psychological safety that make it effective — is one of the highest-leverage actions a safety leader can take.

Conclusion

If the same types of incidents keep happening in your organization, the investigation process itself is the problem. The five mistakes above — stopping at human error, weak corrective actions, confirmation bias, not reaching organizational factors, and investigating in isolation — are systemic issues that require systemic solutions.

The good news: they're all fixable. It starts with recognizing that investigation isn't about finding someone to blame. It's about finding something to fix.