← Article Library
OHS Canada

When the System Stops Listening

How organizations can be data rich, apparently stable, and still lose the ability to hear risk.

Originally published on OHS Canada. This page presents Donald P Andrechek’s original article manuscript.

Most organizations that experience a serious incident share a common reaction afterward.

“We never saw it coming.”

The statement is usually sincere. Leaders review dashboards, audit results, lagging indicators, and assurance reports, and genuinely struggle to understand how a system that appeared stable could fail so suddenly. Investigations then focus on proximate causes, human error, or isolated breakdowns in controls.

What is rarely examined is a more fundamental question: what if the system did see it coming, but stopped listening?

Serious incidents rarely begin with failure. They begin with a gradual breakdown in how truth moves through an organization. Signals are delayed, softened, filtered, or never raised at all. Over time, the system becomes less responsive to its own warning signs, even as formal safety performance appears strong.

This is not a behavioural problem. It is a system condition.

How systems unintentionally silence themselves

Most safety systems are designed with good intent. Reporting processes, incident classifications, escalation thresholds, and performance metrics are meant to create visibility and control. But as systems mature, they often begin to reward the appearance of stability over the presence of truth.

Near misses that do not result in injury are reclassified as low value. Operational concerns that cannot be easily categorized are deprioritized. Informal warnings are replaced with formal language that removes urgency. Over time, people learn what gets traction and what does not.

The result is not silence. It is selective hearing.

Only signals that fit existing categories, align with acceptable narratives, or avoid disrupting production tend to move upward. Signals that challenge schedules, budgets, or reputations tend to stall at the point of origin.

This is how organizations can be data rich and insight poor at the same time.

The role of tempo in signal failure

As operational tempo increases, the system’s ability to process truth decreases.

Under time pressure, decisions are compressed. Conversations become transactional. Reporting becomes abbreviated. The space required to reflect, challenge assumptions, and ask uncomfortable questions disappears. What remains is execution.

In these conditions, people do not stop reporting because they do not care. They stop reporting because the system no longer has the bandwidth to receive what they are seeing. Signals that require explanation, context, or follow up are perceived as friction.

Over time, the system adapts by filtering itself.

This is why organizations often experience a paradox: safety performance indicators look strongest just before a major event. The system is not safer. It is quieter.

Why leaders are genuinely surprised

When an incident finally occurs, leaders are often accused of ignoring warnings. In many cases, they never received them in a form that conveyed risk.

The signals existed, but they were diluted through layers of interpretation, normalization, and risk acceptance. What began as concern at the front line arrived at leadership as routine variation, acceptable deviation, or noise.

By the time consequences appear, the system has already lost its ability to self correct.

This is not a failure of leadership intent. It is a failure of system design.

Investigations that look downstream

Traditional investigations focus on what failed at the point of harm. Equipment. Procedures. Supervision. Training. These analyses are important, but incomplete.

They rarely ask: when did the system first become aware of this risk, how long did it exist before action was taken, what slowed or blocked that signal, what conditions made silence safer than speaking.

Without answering these questions, organizations fix outcomes rather than causes. Controls are added, training is refreshed, and the system returns to normal operation with the same listening capacity it had before.

This is why similar incidents repeat.

Rebuilding a system that listens

Systems that remain resilient over time share a few characteristics.

They treat weak signals as valuable, not inconvenient. They protect the time and authority required to pause work. They reduce the pressure to translate concern into certainty before it is raised. They measure how fast information moves, not just how often it is recorded.

Most importantly, they understand that silence is not the absence of risk. It is a condition created by the system itself.

Listening is not a soft skill. It is an operational capability.

The real leading indicator

If organizations want a true leading indicator of safety performance, it is not the number of reports submitted or audits completed.

It is how quickly uncomfortable truth can move through the system without being punished, reframed, or delayed.

When truth moves freely, systems adapt. When truth slows, systems fail quietly.

By the time harm occurs, the system has already stopped listening.

Author bio

Don Andrechek is a Canadian health and safety professional with more than 35 years of experience across energy, construction, transportation, education, and public sector operations. His work focuses on system failure, human capacity, operational tempo, and preparedness. Don is the creator of the Truth Tempo Preparation framework, which examines how truth flow, work pace, and readiness interact to predict failure long before incidents occur.

Continue reading

More writing from Donald P Andrechek

Return to the article library to explore safety, leadership, TTP, human capacity, AI, education, healthcare and systems.

Browse all articles