Every major industrial accident is reported as a shock — an event that no one could have foreseen. The pattern is consistent across decades and industries. And it is almost always misleading.
Serious accidents are rarely unpredictable. They are the end of a chain of smaller failures, warnings and near-misses that were noticed, documented and then not acted on. The investigation that follows an incident usually finds, not a single cause, but a cascade — and the cascade was visible long before the event.
The normal accident theory
There is a well-established body of work explaining why complex systems fail, and its central insight is uncomfortable.
In systems that are tightly coupled — where each part affects the next with little slack — and interactively complex — where the interactions are not fully understood by anyone — failures are inevitable. They are not the product of negligence alone; they are the product of complexity itself. The accident is waiting in the design.
This is why the same types of incidents recur across different industries: the structure of the system, not the character of the operators, generates the failure modes.
The near-miss economy
The most important sign of trouble is the near-miss — the event that almost went wrong.
Every serious incident is preceded by near-misses, often many of them, usually recorded and rarely acted on. The near-miss is the system telling you where the boundary is. The organizations that treat near-misses as free information, to be investigated and corrected, are the ones that avoid the incident. The ones that treat them as unimportant are, in effect, choosing not to read the warning.
The economics are perverse: the closer the call, the cheaper the lesson. But the lesson is only captured by organizations that want it.
The normalization of deviance
Another recurring pattern is the gradual acceptance of abnormal conditions.
When a system operates slightly outside its safe envelope and nothing bad happens, the deviation becomes normal. The next deviation is a little larger, and the next larger still. Over time, operating conditions that the designers would never have accepted become routine. The accident occurs when the deviation finally exceeds the margin that the system could absorb.
This is why the question “how did it get this bad?” so often has the answer: one small step at a time.
The maintenance squeeze
Look beneath the technical explanations and a common driver appears: deferred maintenance.
Equipment that is run past its service interval, inspections that are skipped to meet schedules, repairs that are postponed to save money — these are the quiet decisions that accumulate into catastrophe. The savings are real and immediate; the costs are deferred and diffuse. In the accounting that matters, the maintenance was always the cheaper option.
The organizations that fail on maintenance are not the ones without resources; they are the ones whose incentives reward the visible over the preventive.
The voice that gets ignored
In nearly every major incident, there was a person who warned in advance.
The technician who flagged the unusual reading, the worker who reported the unsafe condition, the engineer who questioned the shortcut — the warnings were there, and they were overridden or ignored. The pattern is so consistent that it has a name: the failure of the internal whistleblower. The organization, in effect, silenced the information it needed most.
This is why the health of an organization’s internal warning system — the willingness of people to speak up and the receptiveness of leaders to hear them — is one of the best predictors of its safety record.
The blame reflex
After an incident, the reflex is to find a culprit — a person to blame, a head to roll, a story to tell.
The blame reflex is understandable, and it is dangerous. When an organization responds to an incident by punishing individuals, it guarantees that the next near-miss will be hidden rather than reported. The systemic causes — the design, the incentives, the culture — remain unexamined and unrepaired. The same accident is simply waiting for a new set of circumstances.
The organizations that learn from incidents are the ones that resist the blame reflex and investigate the system rather than the scapegoat.
The honest lesson
The pattern across industrial accidents is the same story told in different settings: complexity, normalization, neglect, ignored warnings and the blame reflex.
None of these are exotic. All of them are observable in ordinary organizations, long before anything serious happens. The companies that avoid the big incident are not the lucky ones; they are the ones that read the pattern and acted while there was still time.
The big accidents are reported as shocks, but they are never really sudden. They are the visible end of a long, visible chain. The question is not whether the pattern exists in your organization. It is whether anyone is looking at it before the light goes on.