New Safety Metrics Owner in 60 Days: Turn Dashboard Numbers Into Operating Decisions

A safety dashboard can stay green while critical exposure grows, especially when the measures reward reporting volume, audit completion, or days without a recordable event more than they reveal how work is actually controlled. A new safety metrics owner has sixty days to change that relationship between numbers and decisions.

This role is not a spreadsheet assignment. It is a transition from collecting information to making evidence useful for operations, because every indicator carries a definition, an incentive, a blind spot, and an implied management response.

What a new safety metrics owner needs to understand before starting

The first responsibility is to resist the pressure to make the dashboard look complete before it becomes trustworthy. A metric can be precise and still be misleading when its denominator changes, its reporting rules are unclear, or the people closest to the work have no reason to believe that raising a concern will improve the situation.

Outcome measures still matter. Recordable injuries, lost-time cases, process events, and occupational health outcomes show what has already happened, which makes them important for accountability and trend review. They do not, however, explain whether a critical barrier is present today or whether a supervisor can recognize a deteriorating condition before harm occurs.

That distinction is central to Andreza Araújo's position in Muito Além do Zero, translated as Far Beyond Zero. A rigid zero target can protect the number instead of protecting the person when leaders treat an outcome as proof of capability rather than one piece of evidence about the system.

The metrics owner therefore needs a written operating question for every measure. What decision should this number change? Who owns that decision? What evidence shows that the response happened? If the answer is unclear, the measure may still be useful for research, but it does not belong in a high-consequence executive review without a clear limitation.

First week: map the measurement system before changing it

Start with listening and document review rather than a dashboard redesign. Collect the monthly executive pack, site scorecards, regulatory submissions, incident summaries, action trackers, audit reports, critical-control reviews, and any local spreadsheets that influence performance conversations.

Build a metric register with one row for each indicator. Record the plain-language definition, formula, unit, population, time window, data source, person who enters the data, person who approves it, intended user, target, escalation threshold, and known limitation. When two functions use the same term differently, keep both definitions visible until the accountable leader resolves the difference.

Interview the data owners and the decision owners separately. The person who enters a near-miss report may know why the number changed, while the executive who reviews the monthly trend may know which decisions the number fails to support. The gap between those perspectives is valuable evidence, because it reveals where a dashboard has become detached from the work.

During the first week, also inspect incentives. Ask whether teams are rewarded for low case counts, high observation counts, closed actions, completed training, or audit scores, and then ask what behavior each reward makes easier. As described in A Ilusão da Conformidade, translated as The Illusion of Compliance, good indicators do not guarantee good practices when the measurement process becomes an end in itself.

Days 8 to 30: convert the register into decision rules

By the second week, select a small set of measures that represent different parts of the risk picture. Include at least one outcome measure, one exposure or work-condition measure, one critical-control measure, one action-quality measure, and one signal about reporting or workforce participation when that signal can be interpreted responsibly.

Do not fill the dashboard with every available field. A measure belongs in the executive view when its movement can prompt a specific decision about resources, staffing, design, supervision, maintenance, procurement, or escalation. If the only response is to ask for another presentation, the metric is probably describing activity rather than control.

Give every selected indicator a response card. The card should state what a green, amber, or red condition means, who reviews it, what evidence is required, how quickly the response begins, and how completion is verified. Thresholds should not be copied from another site without checking whether the underlying exposure, reporting process, and operating context are comparable.

Test the cards with a supervisor and a plant manager using a recent reporting period. Ask each person to explain what they would do next, which authority they need, and what would prove that the condition improved. If the answers conflict, improve the decision rule before improving the visual design.

At day 30, publish a one-page metric charter. It should explain what the dashboard is for, what it cannot prove, how definitions are governed, how data corrections are handled, and how leaders are expected to respond when the evidence is uncomfortable. This is where the role becomes visible as governance rather than reporting support.

Days 31 to 45: connect the dashboard to field evidence

A dashboard becomes credible when its signals can be checked against the conditions that produced them. Choose two or three indicators and trace them from the executive view to the source record, the responsible supervisor, the worksite, and the decision that followed.

Suppose a critical-control score improves. The verification question is not whether the chart moved upward. It is whether the control was present, used under pressure, maintained, and capable of preventing the consequence it was meant to address. A completed inspection may be evidence of activity, while a verified control test may be evidence of protection.

Use this trace to identify false comfort. A fall in reported concerns can mean fewer concerns, lower trust, reduced access, a changed definition, or a reporting channel that no longer works. A rise in reports can indicate worsening conditions, greater exposure, stronger trust, or better visibility. The owner should preserve these possibilities until the field evidence narrows the interpretation.

Andreza Araújo makes this distinction practical in Safety Culture Diagnosis: Learn how to do your own. Measurement is useful when it helps an organization compare perception, behavior, and operating conditions instead of allowing one polished score to stand in for culture.

Days 46 to 60: make leadership review the work, not the score

Use the second month to redesign the review conversation. Put the indicator beside the exposure it represents, the control expected to manage that exposure, the decision owner, and the last verification. This arrangement forces the meeting to move from commentary to action.

Ask leaders to answer four questions. What changed in the work? What evidence supports that reading? Which control or resource needs attention? What will be checked before the next review? These questions are deliberately simple, but they prevent a green status from ending the discussion before anyone understands why it is green.

Separate accountability from blame. A leader should own the decision and the resources attached to it, while the data owner should own the quality and traceability of the measure. When those responsibilities are mixed, people often protect themselves by protecting the number, which leaves the underlying exposure without a clear owner.

Run one review as a pilot and record the decisions, not only the minutes. Afterward, ask whether the dashboard changed what the group prioritized, who received authority, or which control was verified. If nothing changed, the dashboard may be informative, but it has not yet become operational.

After day 60: govern the measures without freezing them

The role does not end when the dashboard is published. Establish a monthly definition review, a quarterly indicator-value review, and an immediate review whenever a serious event, major change, reporting anomaly, or regulatory expectation exposes a weakness in the measurement system.

Retire measures that no longer support a decision. Add a measure only when its purpose, owner, data source, and response are understood. Keep a change log so that leaders can distinguish a real trend from a revised formula, a new reporting population, or a shift in data quality.

Protect the dashboard from two opposite failures. The first is metric inflation, in which every concern becomes a new column until attention is diluted. The second is metric worship, in which a small group of familiar numbers is treated as the complete truth about safety. A disciplined system uses fewer measures with clearer meaning and makes uncertainty visible when the evidence is incomplete.

Common mistakes that weaken the role

The first mistake is starting with colors, charts, and software. Visual polish cannot repair a definition that the workforce does not share or a target that encourages silence.

The second mistake is treating leading indicators as proof of prevention. Training completion, inspections, observations, and closed actions can be useful signals, but they do not prove that a critical exposure was controlled when the work became difficult.

The third mistake is putting every metric on the same page. Executives need a decision view, while specialists may need diagnostic detail. Combining both without hierarchy makes the urgent look ordinary.

The fourth mistake is celebrating green without asking what made it possible. A green indicator can represent strong control, incomplete data, a delayed update, a narrow definition, or a reporting process that has lost credibility. The owner should make that distinction visible rather than provide reassurance by default.

The fifth mistake is hiding the limitation. Every important metric has a boundary. Stating that boundary does not weaken the dashboard. It gives leaders a more accurate basis for deciding what must be checked next.

Resources to deepen the work

For the cultural side of measurement, Safety Culture: From Theory to Practice offers a useful bridge between maturity, leadership behavior, and the routines that make safety expectations credible. For diagnosis, Safety Culture Diagnosis: Learn how to do your own helps the reader examine the distance between what an organization says, what people perceive, and what the operation reinforces.

For the specific danger of outcome worship, return to Muito Além do Zero, which challenges the assumption that a favorable result alone proves a capable system. These resources complement technical standards and local requirements, but they do not replace a site-specific evaluation by qualified professionals who understand the operation and its legal duties.

The 60-day decision

At the end of sixty days, the new safety metrics owner should be able to show more than a refreshed dashboard. The role is ready when each important measure has a definition, an owner, a decision rule, a limitation, and a verification path that reaches the worksite.

The strongest dashboard does not promise certainty. It helps leaders see where the evidence is reliable, where the exposure is changing, and where a decision must be tested before the next review. That is how numbers stop serving the appearance of control and start supporting the control itself.

To connect this role profile with a wider safety culture transformation agenda, explore Andreza Araújo's books and Safety School resources.