How to Run a Safety Culture Onboarding Audit in 30 Days

A new plant manager inherits more than production targets, cost pressure and maintenance backlog. The role also inherits a safety culture whose real shape may be different from the version described in procedures, dashboards and welcome briefings. This guide gives a 30-day safety culture onboarding audit for leaders who need evidence before they set priorities.

The common onboarding pattern is too polite. The new leader hears that safety comes first, reviews the latest indicators, thanks the team and promises visibility. That can create trust, but it can also miss the deeper question: when work becomes difficult, who has authority to slow down, challenge a shortcut, escalate weak controls and protect a worker who raises a concern?

Across 25+ years leading EHS in multinational operations, Andreza Araujo has seen that cultural risk often hides in respectable routines. In Safety Culture: From Theory to Practice, she argues that culture appears in repeated decisions, not in declared values alone. A safety culture onboarding audit should therefore test decisions, evidence and control, because friendliness during the first month does not prove operational discipline.

What you need before starting

Before the audit begins, collect the last 12 months of incident records, near-miss data, corrective-action aging, audit findings, stop-work records, contractor performance, critical-control verification and leadership walk notes. The documents are not the conclusion. They are the first map of where field observation should go.

ISO 45001:2018 expects leadership, worker participation, operational control and management review to work together. For onboarding, that means the plant manager should not treat safety culture as a soft topic separated from controls. The audit must ask whether the system can detect weak signals, make decisions and close risk gaps before harm occurs.

Choose one sponsor, usually the plant manager, one EHS facilitator and 6 to 10 interviewees across shifts, maintenance, operations and contractors. Keep the scope narrow enough to finish in 30 days. The point is not to diagnose everything, but to identify the cultural conditions that could mislead leadership during the first quarter.

Step 1: Define the audit question in operational language

Step 1 is to write one audit question that connects culture with work. A useful question is: can this site stop, adapt and escalate work when the planned controls no longer match reality? That wording prevents the audit from drifting into generic opinions about whether people care about safety.

Andreza Araujo's book A Ilusão da Conformidade, glossed for English readers as The Illusion of Compliance, is relevant because the first audit risk is mistaking documented conformity for lived control. If the plant manager asks only whether procedures exist, the answer will usually be yes. If the leader asks whether procedures still govern difficult work, the conversation becomes more honest.

Write the audit question at the top of every interview note and field observation. When a finding does not answer the question, park it for later. This keeps the 30-day review focused on decision quality rather than becoming a broad commentary on morale, communication or personal style.

Step 2: Read the dashboard for what it cannot show

Step 2 is to review the safety dashboard and name the blind spots before visiting the field. Lagging indicators can show injuries already recorded, while leading indicators may show activity, although neither one proves that serious risk is being controlled during routine work.

The plant manager should compare recordable injuries, first aid cases, near misses, SIF-potential events, overdue actions, repeated audit findings and contractor events. The most useful pattern is not the lowest number. It is the mismatch between where work is risky and where the reporting system is silent.

If near misses are almost absent in a high-activity operation, do not celebrate too early. Silence may mean that workers see nothing, but it may also mean that reporting feels useless, risky or politically expensive. The safety reporting myths that keep workers silent are a practical companion for this step.

Document 3 dashboard questions for field testing. For example, why does one shift report more? Why are corrective actions older in maintenance? Why do contractors have fewer near misses than employees despite higher exposure?

Step 3: Walk the site before accepting the narrative

Step 3 is to observe work before accepting the site narrative. The plant manager should spend time at shift start, peak production, maintenance intervention and contractor interface, because culture is easier to read when competing priorities appear at the same time.

Edgar Schein's culture work separates visible artifacts from deeper assumptions. In a plant, visible artifacts include signage, PPE, boards, meetings and forms. Deeper assumptions appear when a supervisor decides whether a late shipment justifies a shortcut, whether a damaged guard waits until tomorrow or whether an operator can stop a task without being labeled difficult.

During the walk, do not ask broad questions such as whether safety is important. Ask what changed today, what job worries the crew, which control would stop the work if missing and who can approve a change in sequence. Those questions expose authority and control more reliably than speeches about values.

Link the walk with the existing gemba walk safety guide when the leadership team needs a repeatable observation rhythm after onboarding.

Step 4: Interview supervisors about authority under pressure

Step 4 is to interview supervisors separately from senior leaders. Supervisors translate culture into daily decisions, and they usually know where the official rule becomes fragile under downtime, staffing gaps, contractor pressure or urgent customer demand.

Ask each supervisor to describe the last time they stopped or delayed work for safety. Then ask what happened afterward, who supported the decision, whether production pushed back and how the decision was documented. A site where supervisors cannot remember a real stop decision may have a compliance story rather than working authority.

James Reason's Swiss cheese model helps keep the discussion away from blaming one person. The question is not whether a supervisor is brave enough. The question is whether the organization has layers that support the supervisor when a control is missing, a worker is exposed or a plan no longer fits field conditions.

Score each interview against 4 signs: clear stop authority, access to resources, confidence to escalate and evidence of follow-up. If supervisors have authority on paper but no resource path, the culture depends on personal courage instead of system design.

Step 5: Test whether worker voice produces follow-up

Step 5 is to test worker voice through closed-loop examples. A culture that invites people to speak but does not respond damages trust faster than a culture that never asked. The plant manager should therefore review whether concerns move from report to decision, action and feedback.

Pick 10 recent concerns, near misses or informal field issues. For each one, check who received it, how it was evaluated, what action followed, who communicated back to the worker and whether the same issue reappeared. This turns a vague conversation about openness into evidence about response quality.

Andreza Araujo's consulting work across 250+ cultural transformation projects reinforces a practical point: voice does not grow because posters invite it. It grows when workers see that leaders handle uncomfortable information without retaliation, delay or cosmetic closure.

If the site already has a weak signal program, compare this step with the 30-day speak-up follow-up loop. Onboarding should not create a parallel channel. It should test whether the existing channel works.

Step 6: Compare declared controls with operated controls

Step 6 is to select 3 high-consequence tasks and compare declared controls with operated controls. Good candidates include LOTO, confined space, vehicle-pedestrian interface, work at height, chemical transfer, machine intervention or contractor maintenance.

For each task, ask the field team to show the control that prevents the credible worst case. If the worst case is caught-in injury, show isolation and restart prevention. If the worst case is vehicle strike, show physical separation or traffic authority. If the worst case is toxic exposure, show monitoring, ventilation, rescue and escalation.

This is where the audit becomes hard to fake. A procedure may be perfect, although the operated control may be missing, bypassed, poorly understood or dependent on one experienced person. In Sorte ou Capacidade, glossed as Luck or Capability, Andreza Araujo argues that the absence of accidents does not prove capability. The same logic applies to every control that has not been tested.

Record each control as verified, partially verified or not verified. Do not turn partial verification into a pass, because partial controls are where serious events often wait.

Step 7: Map the first 5 cultural risks without diluting them

Step 7 is to turn evidence into the first 5 cultural risks. Keep the list short because a new plant manager who announces 20 priorities in the first month teaches the organization that nothing is truly urgent.

Use risk language rather than personality language. Instead of writing that supervisors lack courage, write that stop-work authority depends on individual confidence because escalation support is inconsistent. Instead of writing that workers do not report, write that feedback after reporting is delayed, which makes weak signals less visible.

The table below separates common symptoms from stronger audit findings.

The stronger wording matters because it points to a decision. It also prevents the plant manager from turning cultural diagnosis into a list of attitudes that nobody can own.

Step 8: Convert findings into a 90-day leadership cadence

Step 8 is to convert the 30-day audit into a 90-day cadence with owners, dates and evidence. The first month is diagnostic. The next quarter has to prove that leadership will change routines, not only publish findings.

Choose 3 to 5 commitments. Examples include weekly review of stop-work decisions, monthly audit of overdue corrective actions, field verification for one critical control per week, contractor-interface review before shutdown work and feedback within 48 hours for every safety concern.

During Andreza Araujo's PepsiCo South America tenure, where the accident ratio fell 50% in six months, the lesson was not that one campaign changed everything. The practical lesson was that disciplined leadership cadence can shift what leaders notice, what supervisors escalate and what the organization treats as non-negotiable.

Connect the cadence to the leadership cadence case if the site needs a longer transformation rhythm after the first quarter. The onboarding audit should become the first operating cycle, not a report that fades after the new leader's introduction period.

Safety culture onboarding audit checklist

The checklist below keeps the 30-day audit practical without reducing it to a form. Use it as a review gate before the plant manager presents findings to the leadership team.

• The audit question connects culture with operational decisions.
• Dashboard blind spots were converted into field questions.
• At least 4 field observations covered different shifts or work conditions.
• Supervisor interviews tested authority, resources, escalation and follow-up.
• Worker voice was tested through real concerns and closure evidence.
• At least 3 high-consequence tasks were checked for operated controls.
• The first 5 cultural risks were written as decision risks, not personality judgments.
• The 90-day cadence includes owners, dates and evidence of completion.

Every month that a new plant manager accepts the culture narrative without field evidence, the site teaches people that the first impression matters more than the real risk path.

Final decision after 30 days

A safety culture onboarding audit works when it gives the new leader a disciplined answer to one question: where does this site lose control when work becomes difficult? The answer should come from dashboards, field walks, supervisor interviews, worker concerns and control proof, not from a polished introduction deck.

If your operation needs help turning onboarding findings into leadership cadence, supervisor authority and cultural transformation, Andreza Araujo and ACS Global Ventures can support the diagnostic and implementation plan through Andreza Araujo.