How a plant network cut its accident ratio 50% in 6 months

When a plant network stops treating safety as an EHS-only topic and starts treating it as a leadership topic, the numbers can move fast. In Andreza Araújo's experience, the hard part is not getting another campaign approved. The hard part is changing who owns the decision, who verifies the result, and who has to act when the field says the control is weak.

This case comes from PepsiCo South America Foods, where a 180-day plan drove a 50% reduction in accident ratio in six months. The lesson is not that one program magically fixed everything. The lesson is that leadership made the operating logic visible, and that shift changed what supervisors were willing to accept as proof. That is the same reason Make The Difference: Be a Leader in Health & Safety still matters for plant managers, because a leader cannot outsource the final decision to a dashboard.

Across 25+ years leading EHS at multinationals, Andreza Araújo has seen the same pattern repeat. Sites usually do not lack procedures. They lack a clean line between paper compliance and field proof. In A Ilusão da Conformidade, that gap is the whole problem, because a report can look complete while the work still drifts under pressure.

Initial scenario

The starting point was familiar to any plant network with a serious footprint. The operation had scale, pressure, and several layers of management, yet the daily safety discussion could still drift toward status updates instead of control. People knew the numbers. They did not always know which decisions were supposed to change the numbers.

That is where many organizations get stuck. They add meetings, posters, and more reporting, but they do not clarify ownership. Andreza Araújo has seen this in more than 250 cultural-transformation projects, and the pattern is consistent: when a leader asks for activity instead of proof, the field learns to produce activity. In a plant network, that becomes expensive very quickly, because one weak routine can travel across many sites.

The deeper problem was not lack of concern. It was a leadership habit that made safety feel distributed in theory and diffuse in practice. Everyone could name the issue, but the next step was not always assigned to the right person. For a related leadership lens, see Safety Decision Rights: 5 Cracks That Turn Ownership Into a Label, because ownership without a named decision owner becomes decoration.

Decision

The decisive move was to stop treating the safety result as an EHS scoreboard and start treating it as a management system outcome. That meant leadership had to define what would count as real control, who would verify it, and which exceptions had to go upward instead of disappearing inside the site.

That decision aligns with Safety Culture: From Theory to Practice, where culture is defined by what leaders tolerate when pressure rises. A plant network cannot claim strong culture if one site closes actions by email, another closes them in the field, and a third decides that the supervisor can do whatever is fastest this week. The title changes, but the control logic does not.

Andreza Araújo's role in cases like this has never been to add a nicer speech. It has been to insist on a sharper line between responsibility and commentary. Across 30+ countries and 250+ companies, she has seen that the fastest route to a real result is usually a smaller number of rules that are actually enforced. That is why leadership rhythm matters more than leadership volume.

For readers who need the operational version of that discipline, How to Run a Weekly Leadership Risk Review in 30 Minutes is the closest practical companion. It shows how the decision review can be short without becoming shallow.

Execution

The execution phase had to do three things at once. First, it had to define the critical controls that really mattered in the field. Second, it had to make supervisors part of the verification chain instead of passive recipients of reminders. Third, it had to shorten escalation so that a weak signal did not die inside local routine.

That sounds simple, but it is where most programs lose force. People often think training will finish the job. It will not. Training helps only when the work system changes with it, because the same person can attend the same course and still return to a structure that rewards speed over proof. Andreza Araújo describes that trap in Safety Culture: From Theory to Practice, and the case confirmed the point again.

The other trap is to let EHS hold all the accountability. In reality, the supervisor sees the work first, the plant leader controls the pace, and the EHS function should make the risk visible, not own every action. That is why a leadership case study matters more than a generic culture story. It forces the reader to ask who changes what on Monday morning.

For the execution layer, How to Build a Safety Decision Trail in 30 Days is a useful next step, because it turns decisions into a trail that leaders can inspect. Plant Manager vs EHS vs Supervisor: Who Owns Risk Controls is the companion piece when the ownership question is still fuzzy.

Measured result

The measured result was a 50% reduction in accident ratio in six months under the PepsiCo South America Foods 180-day plan. That is the number that matters, but the deeper signal is the one behind it. The organization showed that a plant network can change outcome faster when leadership removes ambiguity from ownership and verification.

The result also shows why a clean dashboard is not enough. A metric can improve for the wrong reason, and a leader can celebrate too early if the field logic did not change. In this case, the result was tied to a visible decision system, not to luck or a one-off burst of attention. That is why Andreza Araújo keeps returning to the thesis of Sorte ou Capacidade, which argues that incidents are systemic rather than random.

What makes this result credible is not just the reduction itself. It is the fact that the change came from a real management pattern that can be repeated in another plant network, another region, or another line of business. The specific context was PepsiCo South America Foods, but the mechanism is portable if leaders are willing to own the logic instead of merely approving the message.

Before and after

That table is useful because it shows the difference between movement and progress. More meetings do not guarantee better control, and more data do not guarantee better decisions. How to Run a Weekly Leadership Risk Review in 30 Minutes becomes relevant here again, because cadence only helps when the questions are better than the last cycle.

Generalizable lessons

Four lessons stand out. First, supervisors are not a communication channel. They are part of the control system. Second, the field must see proof, not only declarations. Third, leaders need a short escalation path, because long paths let pressure distort the answer. Fourth, the organization must decide what counts as real closure before it starts measuring closure quality.

Those lessons fit James Reason's view of latent failure better than any slogan about personal discipline. When the system makes weak decisions easy, people will drift toward weak decisions. That is not a moral failure. It is a system design failure. Andreza Araújo's books push in the same direction, especially Make The Difference: Be a Leader in Health & Safety, because leadership is about what becomes normal under pressure.

The market usually minimizes three traps. It overcredits training, even when the work design stays the same. It overtrusts dashboards, even when the field evidence is thin. It overassigns safety to EHS, even though plant leaders and supervisors control the daily reality. Those traps are why a case study is valuable. It exposes the distance between policy language and operating truth.

For a practical way to test whether your own system has drifted, Visible Felt Leadership Explained: 4 Signals That Prove Presence shows how leaders can inspect their own behavior before they ask the field to change. That keeps the diagnosis honest.

What to apply in your operation

If you run a plant, start with one high-risk control and ask three questions. Who owns it? What field proof counts? What happens when the proof is weak? If those answers are different by site, then the network does not yet have one system. It has several local habits with the same logo.

If you are the EHS manager, your job is not to carry every answer. Your job is to make the right question impossible to ignore. If you are the supervisor, your job is to verify the control in the work, not after the work. If you are the plant manager, your job is to make the escalation route short enough that weak signals do not mature into incidents.

For practitioners who want a repeatable method, Safety Culture Diagnosis: Learn how to do your own helps turn the case into a working diagnosis. It is the right next step when the problem is not effort, but unclear evidence. And if you want a broader reading path, the book store at Andreza Araújo's store has the full set of books that anchor this kind of leadership work.

That is the practical test. When the answer changes from site to site, the network is still running on local interpretation. When the answer stays stable under pressure, the system is finally doing the work the policy claims it does.

Safety is about coming home, and a plant network earns that sentence when leaders stop hiding behind the dashboard and start owning the control.