How a 19-country operation changed the monthly safety review before the number moved

The monthly safety review looked healthy on paper. Every country sent the deck on time, the rows were green, and the leadership team had enough charts to believe the system was under control. The field told a different story, because supervisors were still solving local exceptions by instinct instead of by a shared decision rule.

That gap is exactly what Andreza Araujo warns about in A Ilusão da Conformidade. A clean form does not prove a working system, and a translated rule does not prove a common standard if the next supervisor still has to guess what to do.

Initial scenario

In a 19-country consumer-goods operation, the presentation language was consistent, but the language of work was not. Each country could repeat the policy, yet supervisors still made local exceptions because the meeting asked for status, not for a decision.

Across more than 250 cultural transformation projects, Andreza Araujo has seen the same failure pattern. Leaders think they have a communication problem, while they actually have a decision-design problem, and the monthly review keeps hiding that distinction.

James Reason helps explain why the visible failure appears late. The final event is only the part everyone can see, because the weak layers were already drifting long before anyone opened the monthly deck.

The decision

The first change was simple and uncomfortable. The monthly review stopped asking only what happened and started asking what decision the meeting had to produce before the next month.

That shift made the rhythm look much closer to a monthly safety metric review cadence than to a presentation. It also forced the team to check metric freshness, because a stale number can flatter the board while the worksite keeps drifting.

The team also made decision rights visible. If the answer required escalation, the review named the owner, the deadline, and the condition that would reopen the item, so the meeting became a control point instead of a ritual.

Execution

The execution had four moves.

First, the team reduced the review to indicators that could trigger action. Second, each red item had one owner. Third, closure required field proof, not a verbal update. Fourth, the item stayed open until the evidence matched the promise.

That discipline is where dashboard latency becomes visible. A dashboard can be full and still arrive too late to govern work, which is why the team also built a shared glossary through a safety metrics dictionary.

The point was not to make every country identical. The point was to make every country answer the same decision question, even when the local control looked different, because common language without common action is only decoration.

Measured result

In the documented PepsiCo case, the broader program reported a 50% accident reduction in six months. In the 19-country governance work, the more important result was that local reality stopped disappearing inside a clean slide deck.

The number mattered, but only because it came after the review changed. If the monthly meeting had stayed a status ritual, the reduction would have been luck, not management.

Generalizable lessons

Three lessons repeat across safety indicators work. First, a metric that cannot drive a decision is a report, not a control. Second, freshness matters as much as precision. Third, one green row does not prove the system is healthy if the field keeps improvising.

• The review should name the next action, not just the past result.
• Every red item should have one owner, one deadline, and one proof condition.
• Local variation is acceptable only when the decision standard stays the same.
• Stale data deserves less confidence than ugly data that arrived on time.

Andreza Araujo makes the same point in Safety Culture: From Theory to Practice and A Ilusão da Conformidade. The work that survives pressure is the work that has been made visible, owned, and verified before the month ends.

What to apply in your operation

If you run a monthly safety review, start with one change. Replace one status question with one decision question, then make the answer carry an owner, a deadline, and a field proof condition.

For leaders who want a practical rebuild, compare your current meeting with metric review cadence and dashboard latency. If the deck still tells a story without changing the work, the review is not finished.

Headline Podcast keeps this kind of case close to the work, not the slogan. If your own monthly review still reports before it decides, rebuild it now and make the next meeting earn its place.