How a 180-Day Safety Reset Made Supervisor Decisions Visible at the Point of Work

When an accident ratio stops improving, the first response is often another training campaign. During Andreza Araujo's PepsiCo South America tenure, the accident ratio fell 50% in six months, but the important lesson was not that people needed more reminders. The result came from changing how leaders and supervisors made safety decisions during real work.

A 180-day safety reset created a short management cycle in which field conditions, production pressure, corrective actions, and leadership decisions could not remain separate. The transferable lesson is a decision rhythm that shows who saw the risk, who acted, what constraint remained, and whether the control worked after the meeting ended.

Initial scenario: the accident ratio needed an operating response

A regional food operation cannot manage safety from a central office alone. Factories, distribution centers, shifts, contractors, and local production decisions recreate exposure every day, which means that a message approved by senior leadership can lose meaning before the next handover.

The initial problem was larger than an injury count. Leaders needed to understand where production pressure changed the task, where supervisors lacked authority to interrupt work, and where maintenance or planning constraints made the expected control difficult to use. A number signals concern, but it does not identify the decision that must change.

Andreza Araujo's experience across multinational EHS roles points to a recurring failure. Leaders approve a safety objective, then delegate uncomfortable choices to a function that does not control schedule, staffing, equipment condition, or budget. The reset moved those choices back into the operating routine.

Decision: the 180-day plan made safety part of the operating cadence

The decisive move was to treat six months as a management cycle rather than a communications campaign. The plan needed a clear start, a defined exposure scope, recurring reviews, and evidence that decisions had reached the field.

That structure changed the question from "What message should we repeat?" to "Which condition can this site change before the next shift?" A message belongs to communication. A condition belongs to an owner who can change a design, remove a constraint, revise a sequence, or stop a task until the control is restored.

The approach is consistent with Safety Culture: From Theory to Practice, where culture is read through repeated management choices. People learn what the organization values when a supervisor can stop work without penalty, a plant manager funds a correction before an incident, and a reported weakness does not disappear into an action register.

Execution: supervisors became the first visible decision layer

Supervisors became the first visible decision layer because they were close enough to see drift and senior enough to interrupt it. Their role was not to patrol for mistakes or repeat a script. It was to compare the expected control with the task as performed, then decide whether work could continue, needed adjustment, or required escalation.

A supervisor who sees a bypassed guard, an incomplete handover, or a congested pedestrian route needs defined authority, a clear escalation path, and a leader who responds consistently when production is affected. Without those conditions, stop-work language becomes ceremonial.

The routine also needed better questions. "Are you working safely?" usually produces agreement because it asks for an attitude statement. "What changed since the plan was written?" tests whether the supervisor understands the task, the control, and the conditions that could make it fail.

Organizations can adapt a safety coaching routine for new supervisors, provided the conversation ends with a decision and not only an observation record.

Evidence: field verification separated activity from protection

A reset can generate impressive activity while leaving exposure unchanged. Meetings happen, forms are completed, and training attendance rises, yet the barrier that matters remains unavailable, misunderstood, or difficult to use during the busiest part of the shift.

Field verification created a different test. Leaders had to see whether the control existed at the point of work, whether the person performing the task could explain its purpose, and whether surrounding conditions supported the expected sequence. The review included equipment, layout, handover, work instruction, and decision authority.

This distinction prevents confusion between closure and control. A corrective action can be closed in software while exposure remains because the spare part is missing, the design is impractical, or the procedure does not match the job. Verification asks whether the condition changed after closure.

A useful dashboard should connect exposure-based safety metrics to decisions about critical controls, repeat deviations, and unresolved constraints. It should not allow a favorable injury trend to hide a deteriorating barrier.

Leadership: production pressure became discussable

Production pressure is not an excuse for weak control, but it is a condition leaders must understand if they want the result to last. When schedule, staffing, maintenance backlog, or changeover sequence makes the safe method difficult, people will eventually improvise unless the organization changes the constraint.

The reset made that pressure visible in leadership conversations. A supervisor could explain why a task was drifting, a site leader could identify the support required, and an executive could see whether the problem needed capital, planning, engineering, or a change in performance expectations.

The question is not whether leaders say that safety comes first. It is what happens when safety and output compete in the same hour. If the only response is to ask the worker to be more careful, the system has preserved the source of the exposure.

Leaders can compare their routines with the contradictions that make field exceptions look acceptable. The goal is to distinguish a controlled temporary decision from a repeated workaround that the operating system has quietly accepted.

Measured result: the 50% reduction was a management signal

The documented result was a 50% reduction in the accident ratio over six months during Andreza Araujo's PepsiCo South America tenure. The short time horizon revealed whether the management system could sustain focus beyond a launch event.

The number should still be interpreted carefully. A lower accident ratio can reflect genuine control improvement, changes in exposure, reporting quality, or work mix. The result is best treated as a management signal that must be tested against field evidence rather than proof that every serious risk has disappeared.

The stronger questions are what became more visible while the ratio changed. Were supervisors escalating earlier? Were repeated deviations receiving process fixes? Were critical controls easier to use? Could leaders identify the owner of an unresolved risk without starting another investigation?

Transfer: smaller operations should copy the decision architecture

A smaller plant does not need a regional bureaucracy to apply the case. It needs a narrow exposure scope, a weekly review, and a disciplined way to connect field evidence to the person who can make the next decision.

Start with two or three serious exposures. Name the owner for each control, define what verification looks like, and agree on the escalation point before the first review. Ask supervisors to bring one changed condition, one control that held, and one constraint that requires leadership action.

Use the first 30 days to establish the baseline, the next 60 days to test control changes, and the final 90 days to examine whether the routine survives production pressure, turnover, and competing priorities. The timeline is a structure for attention, not a promise that every operation will reproduce the same result.

The common error is to copy the target without copying decision rights. The next is to celebrate the percentage while hiding the work required to change conditions. A practical next step is to compare the reset with a stop-work escalation routine.

What leaders should preserve after the reset

The value of a 180-day plan is not the calendar. It is the habit of making risk visible before the injury, giving supervisors authority that works under pressure, and requiring evidence after an action is closed.

Leaders should preserve the weekly decision rhythm, the field verification standard, and the expectation that constraints belong in the operating conversation. They should also keep testing whether workers can report a weak control without being treated as the problem, which connects the case to psychological safety after an incident.

The PepsiCo South America case shows that a 50% accident-ratio reduction can emerge when safety stops being a parallel campaign and becomes a visible management practice. The result is strongest when leaders can explain not only that the number changed, but which decisions changed with it.

FAQ

What happened in the PepsiCo South America safety case? During Andreza Araujo's PepsiCo South America tenure, the accident ratio fell 50% in six months under a 180-day plan that connected leadership cadence, supervisor decisions, field verification, and indicator discipline.

Was the result caused by more safety training? Training was not the central explanation. The transferable lesson was the operating rhythm that made risk visible, clarified who could decide, and required leaders to verify controls during production.

What did supervisors own? Supervisors owned the first practical response to field drift. They observed the task, checked whether the control remained usable, interrupted unsafe conditions, and escalated constraints.

Can a smaller plant use the same approach? Yes. A smaller plant can choose a few serious exposures, establish a weekly decision review, assign owners, and verify changes without recreating a regional program.

Does a 50% reduction prove that safety performance is under control? No. Leaders should also examine serious-risk controls, repeat deviations, reporting quality, and whether the work environment makes safe decisions practical.

The case is not a shortcut to accident reduction. It is evidence that leadership cadence, supervisor authority, and field verification can turn a safety objective into decisions that change the work.