How 250+ Safety Culture Projects Turned Speaking Up Into a Management Control

A worker raises a concern during a pre-job conversation, and the supervisor thanks them before returning to the schedule. The meeting looks respectful. The exposure remains.

Across more than 250 safety culture projects, Andreza Araujo has seen why that pattern persists. Organizations often create channels for employee voice before they create the decision discipline needed to act on it. The result is a polite reporting system that collects weak signals without converting them into safer work.

This case study follows the transformation from voice as a cultural aspiration to voice as a management control. It does not present one company's injury result as a universal benchmark. It describes the recurring operating sequence that Andreza Araujo has used to examine whether speaking up changes the conditions that create risk.

Initial scenario: people could speak, but the work did not change

The initial scenario was not always silence. In many operations, workers already had toolbox talks, suggestion channels, observation cards, open-door policies, or stop-work language. The weak point appeared after the concern was voiced, when nobody could identify who had to decide, how quickly the response should return, or what evidence would show that the exposure had been reduced.

That gap teaches employees an unwritten lesson. Reporting is acceptable when it does not interfere with production, while a concern that requires maintenance, staffing, redesign, or a delayed restart becomes a test of loyalty. The organization may still describe itself as open, although the operating culture rewards quiet adaptation.

Andreza Araujo's work across more than 25 years of multinational EHS leadership places this problem in a wider context. A survey can show whether people feel heard, but the field reveals whether the organization changes a permit, removes an obstruction, repairs equipment, changes a sequence, or accepts residual risk with a named authority.

Decision: define voice as a route to a decision

The turning point is a leadership decision to stop treating participation as the final outcome. A reported concern must have a route that leads to an operational choice. The choice may be to continue, pause, adapt, escalate, or stop, but it cannot remain in a queue whose status nobody can explain.

The decision architecture separates four questions that are often mixed together. What was observed? Who could be exposed? Which control is uncertain or missing? Who has the authority and resources to decide what happens next? This structure prevents a supervisor from reducing a technical concern to an opinion about attitude.

James Reason's work on active and latent failures supports this distinction because the visible action may be only one layer in a wider set of conditions. A worker who raises a concern is providing evidence about the system, not asking leaders to accept a conclusion without examination.

Execution: make the response visible at the point of work

The new process begins with a short response conversation. The supervisor asks the person to describe the condition in observable terms, identifies the possible exposure, and checks which barrier is no longer dependable. The purpose is not to interrogate the reporter. It is to make the decision precise enough for action.

When the control is uncertain, the supervisor names the immediate boundary. Work may pause until a guard is restored, a lifting path is cleared, an isolation is confirmed, or a competent decision-maker reviews the change. The boundary must fit the exposure, because a vague promise to “look into it” leaves the worker carrying the unresolved risk.

The next requirement is ownership. A concern that belongs to everyone belongs to nobody, especially when its solution crosses production, maintenance, engineering, procurement, and EHS. The owner does not need to perform every action, but the owner must be able to coordinate the response and return a decision to the people doing the work.

This is where the approach differs from a recognition campaign. Leaders still thank employees for raising concerns, although appreciation is only the first sentence. The meaningful response is the visible change in the task, the resource allocation, or the operating rule that follows.

Measured result: replace report volume with response evidence

Report volume is a weak standalone measure. A rise can indicate stronger trust, worsening conditions, improved access to a channel, or a mixture of all three. The portfolio lesson is to review the mechanism between voice and control instead of declaring success from participation numbers.

A useful review asks whether the concern reached the correct authority, whether the decision returned within the agreed boundary, whether the affected control changed, and whether the change held after the shift, handover, or production pressure that originally exposed the weakness. Repeated concerns should be grouped by work condition so leaders can see whether the system is resolving causes or only processing symptoms.

The evidence can be reviewed beside injury outcomes, but it should not be confused with them. A safety culture may improve before a lagging indicator moves, and an injury rate may fall temporarily while the reporting system becomes quieter. The measurement task is to connect voice, decision, control, and verification without pretending that one number explains the whole system.

Generalizable lessons from the portfolio

The first lesson is that psychological safety depends on managerial follow-through. People learn whether it is safe to speak by watching what happens to the concern, particularly when the answer requires time, money, or a production compromise.

The second lesson is that standards and voice are not opposites. A worker can question a method because the method is not protecting the intended control, while a leader can require evidence before changing it. Respectful challenge strengthens technical discipline when the response remains specific.

The third lesson is that repeated concerns are governance information. When different crews report the same exposure, the organization should examine design, staffing, maintenance, planning, or supervision rather than asking each worker to be more careful.

The fourth lesson is that a response must be explainable. Even when leaders do not adopt the proposed remedy, they should state what evidence was considered, who owns the remaining risk, and what would trigger a different decision. Silence turns a difficult decision into a cultural penalty.

What to apply in your operation this month

Select one work setting where employees already raise concerns but the response is inconsistent. Review the last five cases and map the time from report to decision, the decision owner, the control affected, and the evidence collected after the work resumed. Do not begin by designing a new platform. Begin by finding where the existing route loses ownership.

Then set one response standard for supervisors. The standard should require an observable description, an exposure statement, a control decision, a named owner, and a follow-up check. Train the supervisors with real cases from the operation, because abstract scenarios rarely reveal the pressure that makes a weak response attractive.

For a deeper operating model, compare this approach with the shift-team questioning routine and the executive decisions that keep risk visible. Both are useful only when the conversation ends in a decision that the field can see.

Conclusion: voice becomes protective when leaders close the loop

The central result from more than 250 safety culture projects is not that every concern should produce the same answer. It is that every credible concern should produce a disciplined response. The organization must know what changed, which control is uncertain, who decides, and how the result will be verified.

Psychological safety becomes operational when a person can surface risk without being punished for making it visible, while leaders remain responsible for testing the evidence and changing the work. Speaking up is then more than a cultural signal. It becomes an early management control that helps the organization act before weak conditions become injuries.

FAQ

Why is speaking up not enough to improve safety?

Speaking up creates information, but information does not protect anyone until a responsible role makes a decision and verifies the control in the work.

How can leaders avoid turning voice into a popularity exercise?

They can ask for observable evidence, define the exposure, assign decision rights, and review whether the control changed instead of rewarding the number of reports alone.

What should happen when a concern is not accepted?

The leader should explain the evidence and reasoning, name the remaining risk owner, and state what new condition would require the decision to be revisited.

Can this model work in a small operation?

Yes. A small operation can use a simple conversation and a visible decision log, provided the owner, response boundary, and verification check are clear.

What is the first action for a safety manager?

Review recent concerns that crossed functions or remained open, then identify the point where ownership disappeared and repair that step before adding another reporting channel.