Safety Culture: 4 Contradictions That Make Field Exceptions Look Like Flexibility

A supervisor notices that a guard is difficult to reinstall during a short maintenance window. The crew develops a faster sequence, finishes the job, and receives praise for keeping the line on schedule. The same workaround appears the following week, then becomes something people describe as practical flexibility.

That is where safety culture becomes visible. The question is not whether workers can adapt. They must adapt when conditions change. The question is whether leaders treat the recurring exception as information that should improve the work or as a convenient replacement for the control that the work was supposed to provide.

Across more than 250 cultural transformation projects, Andreza Araujo has seen that the most revealing culture evidence often sits in this gap between what the organization says it values and what its decisions make easier. The four contradictions below help leaders examine that gap without reducing culture to a slogan or a survey score.

Why recurring exceptions are culture evidence

An exception is a departure from the agreed method, control, or decision boundary. Some exceptions are necessary because a plan cannot anticipate every condition. A field team may need to stop, re-sequence a task, or select a safer method when weather, equipment, access, or simultaneous work changes the exposure.

The cultural signal appears after the deviation. Does the team record what changed? Does a supervisor understand why the original method failed? Does the organization restore or redesign the control? If the same exception returns without review, the operation has created an informal rule that is stronger than the written one.

James Reason’s work on latent conditions is useful here because a visible shortcut may be the final expression of a design, planning, supervision, or resource problem. Holding the operator accountable for the immediate choice still matters, but stopping the analysis there leaves the conditions that shaped the choice untouched.

Contradiction 1: “We value stop-work authority,” but speed earns the reward

Most organizations say that anyone can stop a task when a control is missing or conditions change. The statement sounds credible until the first production delay. If the person who raises the concern is questioned about lost output, while the team that quietly works around the issue is praised for recovering the schedule, employees learn which value has practical priority.

This contradiction persists because leaders often measure the visible result rather than the quality of the decision. A completed order, restarted line, or recovered shipment is easy to see. The risk avoided by a pause is harder to count, so the organization may reward the behavior that creates the most immediate operational relief.

Review the last five stop-work events and compare the response time, escalation path, production consequence, and control restored. If each event becomes an individual negotiation, the organization has not created usable stop-work authority. It has created permission that depends on the courage and status of the person speaking.

Leaders can close the gap by defining what happens after a stop. The supervisor should protect the pause, the technical owner should assess the changed exposure, and the manager accountable for the work should fund the correction or formally accept the remaining risk. The decision should leave evidence, not a private memory of who was difficult.

Contradiction 2: “We want reporting,” but bad news becomes a personal problem

An organization may publish a reporting campaign and still make reporting feel unsafe. The contradiction appears when a near miss, control failure, or concern triggers a search for the person who should have noticed it earlier, while the equipment, workload, or planning condition receives little attention.

People do not need a formal punishment policy to learn this lesson. A dismissive meeting, a public reprimand, or a delayed promotion can be enough to turn reporting into a private calculation. The result is not always silence. Workers may report minor events that are easy to close while keeping high-consequence concerns inside the crew.

Leaders should examine the quality of reports, not only their volume. A useful report describes the exposure, the condition that made it possible, the temporary control, and the decision needed to remove the cause. A report that produces no owner, deadline, or field verification is a record of concern rather than a functioning learning mechanism.

Andreza Araujo’s books, including Safety Culture: From Theory to Practice, emphasize the difference between declared communication and operating trust. The practical test is whether people can bring inconvenient information to the person who controls the work and see a proportionate response that improves the system.

Contradiction 3: “We respect the standard,” but local workarounds are called experience

Experienced workers often know where a procedure does not fit the physical task. Their knowledge is valuable, and a rigid insistence on an unusable method can create its own exposure. The contradiction begins when leaders praise local ingenuity but never ask what the workaround is compensating for.

A workaround may indicate that the tool is poorly designed, the sequence ignores access, the permit is detached from the real job, or the schedule makes the planned control impractical. Calling every workaround unsafe would hide those design problems. Calling every workaround experience would hide the new risk created by the deviation.

Use a three-part review whenever a workaround repeats. First, identify the original control and the reason it was bypassed. Second, describe the replacement control and test whether it protects against the same exposure under the actual conditions. Third, set a named owner and date for restoring or redesigning the original method.

The boundary matters because an adaptation without a review path becomes a permanent exception. A strong safety culture does not demand perfect execution of a defective plan. It makes the defect visible and prevents temporary ingenuity from becoming an invisible standard.

Contradiction 4: “We investigate to improve,” but closure means paperwork is complete

Incident reviews and corrective actions can look disciplined while leaving the work unchanged. The contradiction appears when the organization closes an action because a briefing occurred, a procedure was reissued, or a training record was uploaded, even though the condition that allowed the event remains present.

Administrative completion is attractive because it produces a clear status. The action is either open or closed, and the dashboard can show progress. Field effectiveness is harder because it requires someone to return to the task, observe the control, and ask whether the exposure has actually changed.

James Reason’s model supports a broader review of why an event was possible. The investigation should examine equipment, design, supervision, information, workload, and decision timing, then connect each corrective action to the condition it is meant to change. A new instruction cannot compensate for a missing guard, an impossible maintenance sequence, or a staffing model that removes competent supervision.

Set closure criteria that require evidence from the place where the risk exists. The action owner should show that the control is available, used, understood, and effective under representative conditions. If the proof is only a document, the organization has verified the document, not the control.

What leaders should do when flexibility becomes the operating rule

Begin with the exceptions that repeat, not with the most polished policy. Ask where the agreed method is routinely changed, who knows about it, which production condition drives the change, and what consequence follows when someone refuses the shortcut.

Then classify each exception. Some require a better design. Some require clearer authority. Others require a firm boundary because the replacement method cannot control the exposure. The classification should lead to a decision, an owner, a date, and a field check.

A useful review can use five questions in a leadership meeting.

• Which exception has appeared more than once in the last month?
• What condition makes the approved method difficult to execute?
• Who can authorize the temporary method, and what evidence is required?
• What production signal could pressure people to hide the deviation?
• How will the organization prove that the underlying control was restored or redesigned?

The answers reveal more than a culture statement because they show how the organization handles conflict between safety and production. They also turn vague concern into a management decision that can be checked.

How to tell whether the culture has become more credible

Do not judge progress by the disappearance of reported exceptions. A sudden drop may mean that people stopped raising concerns. Look instead for better evidence: exceptions are identified earlier, escalation takes less time, corrective actions remove the conditions that created the deviation, and supervisors can explain which controls are non-negotiable.

Over time, the organization should see fewer repeated exceptions in the same task, clearer ownership when a control fails, and less dependence on personal heroics to keep work safe. Those changes make culture observable. They show that leadership response, not messaging alone, is shaping behavior.

The strongest culture is not the one that claims no one ever needs to adapt. It is the one that notices adaptation, learns why it was needed, and refuses to let a temporary workaround become the permanent way risk is carried.

For a broader view of how safety culture changes across operations, read what 250+ cultural transformation projects reveal about safety culture.

Andreza Araujo helps leaders turn safety culture into measurable decisions, stronger controls, and safer work. Explore more perspectives on Andreza Araujo’s work in safety culture and leadership.