How to Run a Safety KPI Owner Review in 14 Days

A safety KPI without an owner becomes a number that survives by habit. It may still appear on the dashboard every month, but nobody can explain who validates it, which decision it supports, what should happen when it moves, or whether the data came from the same rule used last quarter. This 14-day review gives EHS managers a practical way to assign ownership before safety indicators become decorative reporting.

What to settle before the review starts

The review starts with a hard distinction. A KPI owner is not the person who extracts the spreadsheet. The owner is the person accountable for the definition, data quality, interpretation, escalation threshold, and management action attached to the indicator. When those roles are split without agreement, dashboards look precise while decision quality decays.

Across 25+ years in executive EHS roles, Andreza Araújo has seen that many organizations do not fail because they lack safety data. They fail because the data has no decision owner. The same metric can mean prevention in one site and bureaucracy in another, depending on who checks it and what leaders do with the signal.

As argued in *Safety Culture: From Theory to Practice* (Araújo), visible routines only become culture when they change real behavior. A KPI review follows that same logic. The purpose is not to make the dashboard neater, but to make each number accountable enough to trigger a field decision.

Step 1: List every KPI that appears in leadership meetings

Start by listing every safety KPI shown in daily, weekly, monthly, and board-level meetings. Include lagging indicators, leading indicators, audit scores, corrective action aging, training completion, control verification, observation quality, SIF precursors, exposure denominators, and any local score used by a plant or business unit.

The common mistake is reviewing only the official dashboard. Informal indicators often shape decisions more strongly than the approved set, especially when a plant manager asks for a local spreadsheet before accepting a safety explanation. If a number influences resources, priorities, bonuses, escalation, or operational permission, it belongs in the review.

For each KPI, record its name, meeting where it appears, current formula, data source, reporting frequency, audience, and the person who currently explains it. This first pass should stay factual. Do not fix the indicator yet, because premature cleanup hides the ownership gaps the review needs to expose.

Step 2: Match each KPI to one management decision

In the second step, match each KPI to one management decision. A useful safety KPI should support a decision such as stopping a risky activity, reallocating supervision, checking a critical control, investigating underreporting, changing contractor conditions, or escalating a deteriorating exposure trend.

This is where many dashboards lose their safety value. They report activity without naming the decision that should follow. Training completion, for example, may look positive while field competence remains weak. A corrective action closure rate may improve while recurrence risk remains untouched.

Use the existing safety metric dictionary as the reference point for definitions, then add a decision column to each KPI. If the team cannot name the decision, mark the KPI as informational only. Informational indicators may stay in background reports, but they should not dominate executive safety meetings.

Step 3: Assign one owner and two supporting roles

The third step is to assign one owner and two supporting roles for each KPI. The owner decides the definition and accepts accountability for interpretation. The data steward checks extraction, coding, timeliness, and source-system consistency. The field verifier checks whether the number matches what is happening in the workplace.

One person can hold more than one role in a small site, but the roles should not be confused. An EHS analyst may be excellent at extraction and still lack the authority to decide what an abnormal trend means for operations. A plant manager may own the decision and still need EHS to verify classification quality.

Put names, not departments, into the review table. Department ownership sounds clean, but it fails when the monthly report is challenged. A KPI whose owner is listed as EHS, Operations, HR, or Maintenance has not yet reached accountability. The review should identify who answers when the indicator is wrong, late, disputed, or ignored.

Step 4: Test the definition against three real records

The fourth step is to test every KPI definition against three real records from the last reporting cycle. Pick one normal record, one borderline record, and one disputed record. Ask the owner and data steward to apply the written definition without relying on memory or local custom.

This test quickly exposes metric drift. If two sites classify the same event differently, the KPI is not comparable. If the owner explains the rule differently from the analyst, the dashboard is carrying hidden interpretation. If the definition requires tribal knowledge, the number will fail during turnover or audit pressure.

For leading indicators, connect this step to the existing leading indicator quality audit. A leading indicator has value only when the definition captures a real precursor or control condition. Counting activities with unclear rules gives leaders volume, not risk intelligence.

Step 5: Check the denominator before judging performance

The fifth step is to check the denominator before judging the KPI. Rates, percentages, completion scores, and exposure-normalized indicators can mislead leaders when the denominator changes quietly. A stable numerator can become a false improvement if work hours, task volume, contractor exposure, or verification opportunities changed underneath it.

The trap is treating the rate as more objective than the count. It is not more objective if the denominator is weak. A site may reduce recordable injury rate while contractor hours move outside the calculation. Another site may look worse because it reports short-duration tasks with greater precision than its peers.

Use the exposure hours, task volume, and control verification comparison to decide whether the KPI denominator fits the decision. For SIF-related indicators, task volume or critical control verification may speak more clearly than total work hours, because fatal risk concentrates in specific work, not in average exposure.

Step 6: Set escalation thresholds before the next report

The sixth step is to set escalation thresholds before the next report is issued. Every KPI owner should define what triggers attention, what triggers action, and what triggers leadership escalation. Without thresholds, leaders negotiate meaning after the number appears, which makes the dashboard vulnerable to politics and optimism.

Thresholds do not need to be complex. A green status may mean the indicator is inside expected variation and field checks show no contradiction. An amber status may mean the trend changed, data quality weakened, or a control signal requires supervisor review. A red status should mean the owner has already named a decision, deadline, and accountable leader.

Be careful with clean green dashboards. The article on dashboard blind spots hiding fatal risk explains why a favorable score can conceal serious exposure when the indicator does not track critical controls. A KPI owner review should make green harder to earn, not easier to display.

Step 7: Review underreporting risk openly

The seventh step is to review underreporting risk openly, especially for injury rates, near misses, unsafe conditions, stop-work cases, and concern channels. If an indicator depends on workers reporting bad news, the KPI owner must understand the social conditions that shape reporting behavior.

James Reason's work on latent failures helps here because weak reporting is rarely only an individual choice. Workers report less when they expect blame, delay, silence, retaliation, or pointless paperwork. Leaders then see a clean dashboard and mistake silence for control.

Use the existing article on underreporting distortions in safety metrics as a challenge tool. Ask each owner which pressure could make the KPI look better without risk actually improving. If the answer is uncomfortable, the review is doing its job.

Step 8: Lock the owner review into a monthly rhythm

The eighth step is to lock the owner review into a monthly rhythm. The first 14 days clean up ownership, definitions, denominators, and thresholds. The monthly routine keeps those decisions alive as work changes, systems change, leaders move, and new risks enter the operation.

In more than 250 cultural transformation projects supported by Andreza Araújo's team, one repeated pattern appears because systems decay when ownership becomes ceremonial. A KPI owner review prevents that decay by forcing a short monthly conversation about whether the number still represents the risk it claims to represent.

The monthly rhythm should answer five questions. Did the definition change? Did the denominator change? Did the data source change? Did the field evidence confirm the signal? Did the owner make or escalate a decision? If the answer is no to the last question for several months, the KPI may belong in a reference report rather than the leadership dashboard.

Safety KPI owner review table

The table below gives the review team a simple structure. It is intentionally operational because the goal is to connect each indicator to a person, a rule, and a decision.

Conclusion: make every safety KPI answerable

A safety KPI owner review should make every important indicator answerable. The team should know who owns it, what it means, where the data comes from, which denominator is valid, what threshold changes the conversation, and which decision follows when the number moves.

For EHS leaders who want to connect metrics to field leadership, Andreza Araújo's Safety School and the book *Safety Culture: From Theory to Practice* offer a practical base for turning reports into management behavior. Start with the 14-day owner review, then use Andreza Araújo's resources to strengthen the routines behind each number.