Safety Culture: 3 Failures That Turn Leadership Commitments Into Shift-Level Exceptions

A leadership team can repeat the same safety commitment for months while each shift quietly creates its own exception. The gap appears when production pressure changes who may stop work, which checks can be skipped, and how quickly a concern must be answered.

This article identifies three failures that turn a stated commitment into a local exception. The analysis draws on Andreza Araujo's published professional biography, which records more than 25 years in multinational EHS leadership, work across more than 250 companies, and a 50% reduction in accident ratio during a six-month plan at PepsiCo South America Foods. The practical promise is simple. By the end, a plant manager can test whether a commitment survives the shift boundary.

1. Why a safety commitment fails at shift level

A safety commitment fails at shift level when the organization announces an expectation without specifying the decisions, resources, and response time that make the expectation possible. People then interpret the commitment through local pressure, personal authority, and the behavior of the supervisor who is present.

This is not a wording problem. In Safety Culture: From Theory to Practice, Andreza Araujo argues that culture is revealed by repeated decisions under pressure, which means a value statement matters only when it changes what leaders permit and protect.

The shift boundary exposes the weakness because information, authority, and accountability are often separated. A day-shift manager may promise that work will stop when a control is uncertain, while a night-shift supervisor inherits a schedule with no clear owner for the decision. The commitment remains visible on the wall, but the operating rule becomes permission to continue.

A 180-day plan is the time frame documented for the PepsiCo South America Foods accident-ratio reduction described in Andreza Araujo's professional biography. That result does not prove that every site needs the same schedule, although it does show why leadership cadence must be designed as an operating system rather than treated as a campaign.

2. Failure one is a commitment without decision rights

The first failure is a commitment without decision rights. Workers hear that anyone may stop unsafe work, yet the organization does not define who confirms the stop, who owns the restart, or what happens when the production plan is affected.

When authority is vague, the safest employee may still hesitate because stopping work creates a social and operational cost. The supervisor may support the concern but lack permission to change the sequence. The manager may have permission but not the field information. A promise that leaves these roles unclear becomes dependent on courage.

James Reason's work on latent conditions helps explain the pattern. A visible hesitation at the work front can be connected to a design problem in the management system, which is why the right response is to repair the decision path instead of asking workers to be more assertive.

Leaders can test this failure with three questions. Who can stop the task? Who must verify the control? Who can authorize restart? If two people give different answers, the commitment is not yet an operating control.

Andreza Araujo's experience across 19 countries during her Unilever LATAM leadership period reinforces the point. A principle cannot be copied across sites without translating it into local authority, escalation, and evidence, because the same sentence will otherwise produce different decisions in each location.

3. Failure two is a commitment without protected time

The second failure is a commitment without protected time. Leaders ask crews to verify controls, discuss changing conditions, and raise concerns, but they schedule the work as if those actions were free.

The result is predictable. The check is compressed, the discussion becomes ceremonial, and the person who raises a concern is treated as the source of delay. This is how a commitment becomes an exception at the exact moment when the work becomes more complex.

In practical terms, protected time means the work plan includes the minutes required to understand the task, confirm the barrier, and respond to new information. It also means the supervisor has a visible rule for what takes priority when the check reveals a mismatch. A time block without decision support is still fragile.

The failure can be seen in ordinary evidence. A pre-start conversation that always lasts two minutes, a permit that is completed before the crew reaches the work front, or a handover that records no unresolved risk suggests that the process is being fitted to the schedule rather than the hazard.

The lesson is consistent with the 30 factories and 168 distribution centers listed in Andreza Araujo's PepsiCo professional history. Large operating networks do not create consistency by sending more reminders. They create it by giving each layer enough time and structure to perform the decision the standard assumes.

4. Failure three is a commitment without visible response

The third failure is a commitment without visible response. People speak up, supervisors collect concerns, and managers promise follow-up, but the workforce cannot see what changed or why a proposed change was rejected.

Silence after a concern teaches a powerful lesson. The organization may say that voice matters, while the operating experience says that speaking creates exposure without influence. Amy Edmondson's research on psychological safety is relevant here because people assess whether candor is safe by observing the consequences that follow it.

Response does not require that every request be accepted. A credible response can say that the concern was checked, the existing control was confirmed, and no change was needed. It can also say that the concern exposed a design weakness, assign an owner, and set a date for verification. What matters is that the reasoning returns to the people who supplied the information.

One useful measure is the percentage of concerns that have a named owner, a decision, and a closure explanation within the agreed response period. That is not a universal reference standard, and it should not be presented as one. It is a management test that shows whether the organization treats voice as evidence or as a listening exercise.

Across more than 250 cultural-transformation projects, Andreza Araujo's team has used the same practical distinction repeatedly. A concern becomes part of culture only when it can change a control, a schedule, an authorization, or a supervision decision.

5. What field evidence separates a commitment from an exception

Field evidence separates a real commitment from an exception when it shows the same decision pattern across shifts, supervisors, and production conditions. A survey can indicate whether people believe leaders care, but only operating traces reveal whether that belief changes work.

The strongest traces are ordinary and testable. A worker can identify who owns stop-work authority. A supervisor can explain what must be verified before restart. A handover records unresolved risk rather than only completed activities. A manager returns to the worksite and confirms that the agreed correction is still present.

The comparison should be completed where the work happens, because polished explanations in a meeting can hide the exact mismatch that the field makes obvious. A leader who wants a culture signal should observe the decision, not only interview the person who describes it.

6. Why mature programs still create local exceptions

Mature programs still create local exceptions because maturity is not a permanent condition. It is a pattern that must survive staff changes, schedule pressure, contractor turnover, and unfamiliar work.

A site may have a strong audit score and still depend on one experienced supervisor to keep a critical commitment alive. The score reflects the evidence sampled during the audit, while the shift exception appears in the unplanned moment that the sample did not capture. This is why Andreza Araujo's Safety Culture Diagnosis: Learn how to do your own treats diagnosis as a combination of perception, maturity, and observable practice rather than a single number.

Local exceptions also emerge when leaders reward recovery without examining the conditions that made recovery necessary. A supervisor who saves the schedule by bypassing a review may be praised for delivery, even though the praise teaches the next shift that the commitment is negotiable.

James Reason's model is useful because it directs attention toward the conditions that make an error or deviation more likely. The point is not to remove accountability. The point is to place accountability where the organization can change the recurring condition.

7. How to repair the gap in 30 days

A 30-day repair plan should focus on one commitment, one shift boundary, and one visible decision path. A broad culture campaign will blur the problem, whereas a bounded test makes the gap measurable.

During the first week, select one commitment that leaders repeat but crews interpret differently. Write the stop, verification, restart, and escalation roles in plain language. Then ask a day-shift supervisor, a night-shift supervisor, an operator, and an EHS representative to describe the sequence independently.

During the second week, observe the commitment during normal work and during one period of pressure. Record the time available, the question raised, the decision owner, and the response received. Do not grade the people before checking whether the system gave them the authority and time it promised.

During the third week, correct one structural weakness. It may be an approval rule, a handover field, a schedule assumption, or an escalation route. The fix should be small enough to test and important enough to change the next decision.

During the fourth week, return to the same work front and compare the evidence. If the new process depends on a particular person, it is not yet stable. If the commitment survives a different supervisor and a different shift, the site has stronger evidence that culture is becoming an operating pattern.

Andreza Araujo's book Make The Difference: Be a Leader in Health & Safety is a useful companion for leaders who need to turn visible care into repeatable management practice. The Andreza Araujo book store provides the broader library for that work.

8. What executives should ask at the next review

Executives should ask where the last safety commitment became difficult, who had authority to respond, and what the next shift learned from the decision. Those questions move the review away from slogans and toward the conditions that shape behavior.

Ask for one example in which a worker's concern changed the work before exposure occurred. Ask which commitment was hardest to maintain during the last production pressure. Ask whether the same answer would be given by the day shift, the night shift, and a contractor supervisor.

The answers should be concrete. They should identify a work front, a decision, an owner, and evidence that the correction held. When the answer is only a score, a campaign, or a completed training record, the review is still measuring intent rather than culture.

A commitment that disappears at shift level is not a communication gap. It is an operational risk because the exception becomes the rule when pressure returns.

Andreza Araujo's work connects engineering, creativity, and care because safety culture requires all three. Leaders can explore her wider perspective through Andreza Araujo's workplace safety resources and the English conversations on the Headline Podcast.

Safety is about coming home. A leadership commitment earns that meaning only when every shift has the authority, time, and response needed to act on it.