New Safety Data Owner in 90 Days: 4 Decisions for a Trustworthy Dashboard

A new safety data owner has 90 days to turn scattered reports into evidence that changes decisions. The role is not a reporting clerk. It connects definitions, field signals, control verification, and leadership action so a green dashboard cannot hide an unresolved serious exposure.

The transition is difficult because safety data usually belongs to several people at once. EHS owns the method, operations owns the work, supervisors create much of the source information, and executives decide what receives attention. When nobody owns the chain from observation to decision, the organization produces polished numbers that cannot explain what changed in the work.

This role profile gives a newly appointed safety data owner a practical 90-day sequence. It builds on the indicator testing method, complements the safety evidence review routine, and keeps the monthly dashboard connected to executive decisions rather than presentation quality.

What a safety data owner needs to understand before starting

A safety data owner is responsible for making safety information trustworthy enough to support action. That responsibility includes definitions, source quality, interpretation, review cadence, and the visible connection between a signal and a decision.

The role does not replace the EHS manager, the operations leader, or the supervisor who controls the task. Instead, it creates a reliable handoff among them. The owner asks whether an indicator has a clear purpose, whether the source can be checked, whether the number reflects exposure, and whether someone has authority to respond when the signal worsens.

ISO 45001:2018 requires organizations to monitor, measure, analyze, and evaluate occupational health and safety performance, while the ISO standard describes the management-system context in which that evaluation occurs. The requirement does not make every metric meaningful. The new owner must decide which measures answer a live management question and which only document activity.

Andreza Araujo's book Safety Culture: From Theory to Practice is useful here because it links maturity to what people repeatedly do, not only to what a system records. A dashboard that counts observations without showing whether controls improved may be complete as a database and weak as a safety instrument.

First week: map the decision chain before changing the dashboard

During the first week, map how a safety signal moves from the field to a decision owner. Start with five recent examples, such as a near miss, a critical-control verification, a late corrective action, a serious-risk escalation, and a recurring exposure that appears in more than one area.

For each example, record who created the source entry, who reviewed it, who interpreted it, who could change the work, and what decision followed. OSHA's reporting requirements show why definitions and timelines matter when an event crosses a reporting threshold, but internal data also needs local decision rules that explain what happens before a regulatory deadline.

Do not begin by redesigning colors, charts, or dashboard software. The common early mistake is to improve the display before understanding whether the underlying data has a stable definition. If one site counts a completed action when it is assigned and another counts it when the control is verified, the comparison is visually attractive and analytically false.

First 30 days: establish four decisions that protect data quality

By day 30, the new owner should document four decisions that make the reporting system dependable. These decisions create the operating boundary for the role and prevent the dashboard from becoming an unowned collection of numbers.

1. Decide what each indicator is meant to change

Every priority indicator needs a decision statement. “Review near misses monthly” is not enough because it does not say what leaders will do with the information. A stronger statement identifies the possible response, such as changing a control, assigning an engineering action, revising a work sequence, or escalating an unresolved exposure.

Four fields, purpose, decision owner, review cadence, and escalation threshold, should appear beside every priority measure, because a value without a response path is only a description.

2. Decide who owns the source, not only the report

The person who publishes the dashboard may not be the person who can validate the source. Assign a source owner for each measure and define what evidence that person must preserve. For a critical-control verification rate, the evidence may be a field record, the control status, the verifier, and the date of the check.

HSE's guidance on permit-to-work systems illustrates the same principle in a different setting: a formal record has value when it represents a real authorization and control process, not when it merely contains the required fields. Data ownership should work the same way.

3. Decide how missing data will be treated

Missing data should remain visible instead of being converted into zero, complete, or not applicable without justification. The owner should create a small data dictionary that distinguishes no exposure, no report, not reviewed, not applicable, and unknown. Those states have different meanings and should not be collapsed for convenience.

Use one example from each major source system to test the definitions. If a supervisor cannot tell whether an overdue action is delayed, closed, or awaiting evidence, the dashboard needs a workflow correction before it needs another chart.

4. Decide when a signal must change the meeting

A metric becomes operational when its movement changes the agenda, the decision owner, the resource request, or the field verification plan. Define the trigger before the next reporting cycle so the team does not negotiate the meaning of a worsening value after seeing it.

Days 31 to 60: test whether the numbers describe the work

During the second month, compare dashboard results with the conditions that generated them. Select three indicators with different behaviors, such as a leading activity measure, a corrective-action measure, and a serious-risk measure, then trace each one back to the field.

The test should include the source record, the person who entered it, the supervisor who reviewed it, and the decision that followed. Three comparisons, reported activity versus verified control, closure versus sustained change, and trend versus exposure, reveal more than a single monthly percentage.

Use NIOSH research and guidance as an external reference point when the indicator concerns exposure, injury prevention, or occupational health. The source does not replace local validation. It helps the owner distinguish a measure that reflects a recognized prevention question from one that was selected only because the system can export it.

At this stage, Andreza Araujo's experience across 25+ years of multinational EHS leadership provides a useful brand anchor, but it should not become a substitute for evidence. The point is practical: safety data earns authority when leaders can trace it to a condition, a decision, and a change that someone can verify.

Month 3: build the review routine leaders can actually use

By day 90, the owner should run a review that takes leaders from signal to decision without requiring a separate explanation for every chart. Put the highest-consequence questions first, then show the evidence needed to answer them.

A useful sequence starts with unresolved serious exposures, moves to critical-control status, checks whether corrective actions changed the work, and ends with measures of participation, reporting, or learning. This order prevents a high volume of low-consequence activity from burying a small number of conditions that could cause severe harm.

For each priority item, the dashboard should show the current state, the previous state, the evidence behind the change, the accountable decision owner, and the next verification date. If the owner cannot provide one of those fields, the limitation belongs in the report instead of being hidden.

James Reason's work on latent failures supports this discipline because organizational accidents can develop through conditions that remain distant from the final event. A dashboard that reports only the last action misses the weakened barriers that made the exposure possible. A good review therefore asks which condition is still present even after the immediate task has ended.

Common mistakes in the first 90 days

The first mistake is becoming the only person who understands the data. That creates dependence, slows decisions, and makes the system fragile when the owner is absent. Publish definitions and decision rules where source owners can use them.

The second mistake is treating completeness as reliability. A fully populated record can still describe the wrong thing, use inconsistent boundaries, or omit the evidence that would test the claim. Reliability comes from a definition that survives comparison with the work.

The third mistake is measuring activity without checking consequence. A rise in observations, conversations, or completed actions may reflect stronger reporting, a campaign, a change in staffing, or a new recording habit. Interpret the number alongside exposure and control evidence.

The fourth mistake is waiting for the monthly meeting to escalate a serious concern. A reporting cadence should never override the risk response required by the condition. When the evidence indicates immediate danger, the escalation route should operate immediately and the dashboard should record what happened afterward.

Resources to deepen the role

Read Make The Difference: Be a Leader in Health & Safety when the data owner needs to influence supervisors and line managers rather than operate as a back-office analyst. Use Safety Culture: From Theory to Practice when the challenge is the gap between declared values and repeated decisions.

Andreza Araujo's practical work also connects measurement with culture diagnosis, visible leadership, and the quality of conversations around risk. The new owner should use those ideas to improve the decision system, not to decorate the dashboard with brand language.

Every reporting cycle that leaves ownership unclear increases the chance that a weak signal will be summarized, filed, and forgotten before the person with authority sees it.

A new safety data owner succeeds when the organization can answer four questions without hesitation: what does this signal mean, who owns the source, what decision does it trigger, and how will the field confirm that the decision worked? The 90-day transition is complete when those answers are visible in the routine, not dependent on one specialist's memory.