How PepsiCo Made a 50% Safety Gain Hold at Shift Level

F5 case study for frontline supervisors, EHS leaders, and operations managers

PepsiCo South America's documented 180-day safety result shows what a 50% reduction in accident ratio can and cannot prove. The result matters because it links a measurable outcome to a defined management period, but the deeper lesson is operational: safety improvement becomes durable only when supervisors can verify the conditions behind the number during ordinary work.

A safety metric can improve before the work system becomes more dependable. That possibility is why a case study should not end with the percentage. It should ask which decisions changed, how leaders checked those decisions at the worksite, and whether the operating rhythm made the safer choice easier to repeat.

During Andreza Araujo's PepsiCo South America tenure, the accident ratio fell 50% in 6 months. This article uses the documented result as a disciplined case lens for supervisors who need to connect a headline outcome with evidence at the point of work.

What the PepsiCo result actually establishes

The PepsiCo result establishes that a 180-day safety plan can be associated with a 50% reduction in accident ratio when the outcome is measured across a defined period. It does not establish that a poster, a single training event, or a lower reporting rate caused the change.

That distinction protects the value of the case. A credible safety result needs a denominator, a time window, and a clear definition of the outcome. It also needs leaders to ask what changed in exposure, supervision, work design, and reporting confidence, because an accident ratio is an outcome measure rather than a complete view of risk.

Andreza Araujo's book Safety Culture: From Theory to Practice frames culture as something visible in daily decisions, not as a slogan that sits above them. The PepsiCo case is useful for the same reason. It invites leaders to examine the decisions beneath the metric.

The Bureau of Labor Statistics publishes injury and illness data by year and industry, giving leaders a public reference for understanding why rates need context rather than celebration alone.

Which decision changed before the number moved

The most important decision in a 180-day reset is deciding what supervisors must verify before work continues. If the plan only asks managers to discuss safety, the organization may improve attention without changing exposure.

A stronger decision names the condition that matters, the person who can check it, and the response required when the condition is absent. For maintenance, that may involve isolation evidence, access control, a current permit, or a rescue arrangement. For material handling, it may involve pedestrian separation, equipment condition, and a route that remains workable during the busiest hour.

OSHA specifies the process-safety elements in 29 CFR 1910.119 because procedures, training, mechanical integrity, and management of change fail when ownership is split across documents and departments.

The Headline Podcast lens adds a practical question. Can a shift supervisor explain, in less than 2 minutes, what evidence proves that today's critical control is available? If the answer depends on a dashboard that updates tomorrow, the decision system is slower than the exposure.

How the 180-day plan becomes a shift routine

A safety plan becomes operational when its expectations fit inside the workday's existing decision points. The supervisor needs a short routine that identifies the exposure, checks the barrier, and records the response when the barrier is weak.

One useful rhythm has 3 moments. Before the task, the supervisor confirms the critical condition and the person who owns it. During the task, the supervisor checks whether the condition remains true after production changes, handoffs, or interruptions. At the end of the shift, the team records what changed and what must be escalated before the next crew inherits the work.

ISO describes ISO 45001:2018 as a management-system framework that requires operational control and attention to change, but the standard does not make a form effective by itself.

When the routine encounters a missing control, the response should be defined before the event occurs. The supervisor may pause the task, apply a temporary measure, call the control owner, or escalate to a manager with authority to change the condition.

What supervisors need to verify at the point of work

Supervisors should verify the condition that prevents harm, not only the activity that signals attention. A completed briefing proves that people met; it does not prove that the worksite is ready.

Four evidence questions make the difference visible. Is the control present? Is it suitable for the task as it is actually unfolding? Is someone authorized to act when it fails? Is there a record that shows the response rather than only the observation?

James Reason's work on latent failures helps explain why a frontline error can be the last visible point in a much longer chain. A supervisor who checks only behavior may miss the design, staffing, equipment, or decision-rights problem that made the error likely.

That is why Andreza Araujo's experience across more than 250 cultural transformation projects should not be reduced to a motivational lesson. The transferable question is whether leaders make the right condition easier to maintain, because asking workers to compensate for weak design eventually turns vigilance into fatigue.

How leaders separate improvement from a better-looking dashboard

Leaders can test whether improvement is real by comparing outcome movement with control evidence, reporting confidence, and exposure conditions. A single green number is not enough.

The test should use at least 4 views. First, review the accident ratio and its calculation period. Second, inspect whether high-risk work changed in frequency, duration, or design. Third, sample closed actions in the field instead of trusting the tracker. Fourth, ask workers what happens after they report a weak control.

A result becomes more credible when the same operating change appears in records, observations, supervisor decisions, and worker accounts. At 30 days, leaders can inspect adoption. At 90 days, they can test whether the routine survives schedule pressure. At 180 days, they can compare the outcome with the evidence trail and decide which controls deserve permanent investment.

50% in 6 months is the documented PepsiCo South America result; the management question is how the organization verified that the improvement represented safer work rather than a narrower record.

What this case teaches beyond PepsiCo

The case teaches that a measurable result is most valuable when it reveals the management system behind it. The number should start a verification conversation, not close one.

Three lessons transfer across industries. First, define the operating condition before choosing the indicator. Second, give the person closest to the work a response path that matches the hazard. Third, review the evidence at a cadence that is faster than the risk can drift.

During her PepsiCo tenure, Andreza Araujo saw a 50% accident-ratio reduction over 6 months, but the defensible takeaway is not that every organization can reproduce the percentage. The defensible takeaway is that leaders need a management rhythm capable of connecting intent, verification, and consequence.

For a related diagnostic, compare the four tests for when tolerance rewrites a safety standard with the field test for a safety rule before an audit. Those articles extend the same question from one case into everyday decisions. For ownership boundaries, read the responsibilities between risk acceptance and action.

What to apply during the next 30 days

Start with one exposure that matters and one shift where the evidence can be observed. Do not launch a 180-day program before the first 30 days have shown that supervisors can identify the control, verify it, respond to failure, and leave a usable record.

During week 1, name the critical condition and its owner. During week 2, observe the routine during normal production rather than during a staged audit. During week 3, sample the response to a weak control and remove any approval step that delays protection. During week 4, compare the field evidence with the metric the leadership team reviews.

Every week in which a safety result is celebrated without checking its operating conditions increases the chance that the organization will preserve the number while losing the control that made the number credible.

Andreza Araujo's work and the Headline Podcast return to the same leadership discipline: make the safe decision specific enough to verify, fast enough to use, and owned by someone who can change the work. That is how a case becomes a practice rather than a story.