After Michael's accident, I sat down with a blank sheet of paper. Not to write a report. Not yet. First, I needed to see the truth. I drew a simple line across the middle of the page. At the right end: Michael's fall. At the left end: everything that was supposed to prevent it. Then I started mapping the defenses that failed: Physical barrier: Anchor points inadequate—known for 18 months Procedural barrier: Risk assessment didn't match reality—never updated after site modification Administrative barrier: Budget approval process—£1,200 modification stuck in queue Supervision barrier: No verification that the job could be done safely—production pressure overrode safety concerns Training barrier: Workers taught to clip on—but not what to do when the system made that impossible Five barriers. All failed. Not because five people made mistakes, but because the system had five holes—and they lined up. This is one of the tools I use in investigation: Barrier Analysis. Not fancy software. Not c

His name was Michael. Thirty-two years old. Two kids under 5 years old. I stood in the plant manager's office reviewing the preliminary incident report. "Fall from height," it read. "Inadequate fall protection. Worker failed to clip on." Three sentences. One life reduced to a procedural violation. But I'd learned to read between the lines. So I went to the site. Talked to Michael's work mates. Not in the meeting room—at the job site, during their breaks, where people actually talk. "We told them the anchor points were too far apart," one said. "You can't reach the next clip point without unclipping from the last one," said another. "Michael knew. We all knew. But the job had to get done." I looked up at the work area. The anchor points were twelve feet apart. The safety harness lanyard was six feet long. The math didn't work. It had never worked. Michael didn't fail to clip on. The system failed Michael. I pulled the maintenance records. The anchor configuration had been identified as

The forklift driver was texting. - He didn't see the pedestrian. - The collision caused serious injuries. Open and shut, right? All the investigators in the room nodded. Write the report. Ban mobile phones, on the shop floor. Add the ban to the induction training. Done. Then someone ask the question, "Why was the pedestrian there?" Silence. The pedestrian was a supervisor. He was taking a shortcut through the warehouse because the walkway added five minutes to his route, between offices. Everyone took the shortcut. They had done for years. "Why isn't anyone using the designated walkway?" Because it was designed when the facility was smaller. The offices were relocated three years ago. But the walkway layout stayed the same. Taking the designated route meant walking three sides of the warehouse instead of cutting straight across the work floor. Why has no one redesigned the walkway? Because the warehouse layout has not been designated as anyone’s responsibility. The operations team assu

The investigation takes five minutes: Worker bypassed safety device: Check Violated procedure: Check Disciplinary action: Check Case closed: Check I've seen this investigation a hundred times. Different industries, different countries, same conclusion. And it's worse than useless—it's dangerous. Because here's what the five-minute investigation didn't ask: Why did he bypass the interlock? Production was behind schedule. His supervisor needed the job finished before shift end. Why was production behind? Two machines down for maintenance, but the production target hadn't changed. Why didn't he use the lockout procedure? Last time he did, it took 40 minutes to get the supervisor to sign off. He'd been told he was "slowing things down." Why was his colleague in the danger zone? They'd developed a workaround—one person bypasses, the other watches. Everyone knew. No one stopped it. It saved time. Suddenly, we're not investigating a "stupid mistake." We're investigating production pressure, r

I've investigated over 500 industrial accidents across four decades. More than 200 of them were fatal. And here's what I've learned: if you're looking for someone to blame, you'll find them. But you won't find the truth. The scaffold that collapsed, the machinery that malfunctioned, the worker who "should have known better" – these are rarely the real story. They're symptoms. The real investigation begins when you ask: What system allowed this to happen? Every accident I investigate tells me two stories. The first is obvious – the immediate cause, the broken procedure, the moment everything went wrong. That's the story most organisations want. It's clean. It's conclusive. Someone gets disciplined, a policy gets updated, and everyone moves on. The second story? That's the uncomfortable one. It's about organisational culture, unspoken pressures, resource decisions made months earlier, and the gap between what the safety manual says and what actually happens on the ground. The question is