Leading Indicators: 6 Gaps That Make Safety Activity Look Preventive

Leading indicators are supposed to give leaders time. They should show whether exposure is growing, whether controls are weakening, and whether the organization is learning before injury confirms the failure. In many companies, however, the leading indicator set becomes another activity scoreboard. It counts what people did, not what the work revealed.

The difference matters because a plant can complete safety talks, inspections, observations, audits, and training while serious risk remains untouched. Activity can rise even as risk intelligence falls. When that happens, the dashboard gives leadership a feeling of prevention without the discipline of prevention.

Across 25+ years in executive EHS roles, Andreza Araujo has seen that organizations rarely lack safety activity. The harder problem is deciding which activities deserve to be treated as early risk signals. This article diagnoses six gaps that make leading indicators look preventive while the operation keeps relying on luck, memory, and informal recovery.

Why leading indicators fail when activity replaces control

A leading indicator fails when it gives leaders confidence without testing the condition that can prevent harm. Training completion, inspection volume, observation counts, safety meeting attendance, and corrective action closure can all be useful. None of them proves prevention by itself, because each one can be performed without changing the risk condition in the field.

James Reason's work on latent failures helps explain the weakness. Harm is often prepared by decisions, design gaps, maintenance drift, production pressure, poor supervision, and weak defenses long before the final event. A leading indicator should help leaders see those conditions while they can still intervene.

In her Portuguese title Muito Alem do Zero ("Far Beyond Zero"), Andreza argues that the absence of accidents can protect the number rather than the life. Weak leading indicators create a similar trap. They protect the appearance of prevention while the organization avoids asking whether the controls that matter are actually working.

Gap 1: The indicator counts participation instead of exposure

The first gap appears when the indicator measures participation but not exposure. A site reports that supervisors completed safety walks, workers attended toolbox talks, and teams submitted observations. Those numbers may show involvement, but they do not show whether people looked at the tasks most likely to produce severe harm.

This is why an observation program can look healthy while missing line of fire, uncontrolled energy, confined space entry, lifting, mobile equipment, or work at height. Participation tells leaders that the routine happened. Exposure tells leaders whether the routine reached the work that can hurt someone tomorrow.

The practical test is simple. Pull a sample of recent leading indicator records and ask which exposure each record describes. If the answer is housekeeping, generic behavior, or a safe condition with no connection to critical work, the indicator is not leading enough for a serious risk conversation.

The article on auditing leading indicator quality gives a method for that sample review. For an F1 diagnostic, the key point is sharper: participation without exposure is a social signal, not a prevention signal.

Gap 2: The control is named but never verified

The second gap appears when a leading indicator names a control but never verifies whether the control is present, effective, and used as designed. The dashboard may say that lockout checks, pre-use inspections, permits, or emergency drills were completed. Completion is not the same as control reliability.

A permit can be signed without isolating the real energy source. A pre-use inspection can be ticked off without challenging a worn sling. A drill can occur without testing the rescue time that matters. When the indicator stops at completion, it may confirm paperwork while leaving the physical or organizational barrier untested.

Andreza Araujo's work in more than 250 cultural transformation projects reinforces a practical lesson: field credibility depends on whether the metric matches what workers see. If workers know a control is weak but the dashboard says the routine was completed, the leading indicator loses moral authority.

EHS managers should separate activity completion from control verification. The existing analysis of exposure hours, task volume, and control verification explains why different denominators answer different questions. For high-risk work, control verification often speaks more clearly than a broad count of completed activities.

Gap 3: The indicator has no threshold that changes a decision

The third gap is a leading indicator with no decision threshold. The number moves, leaders discuss it, and the meeting continues without changing resources, supervision, operating permission, contractor requirements, or escalation. A metric that never changes a decision becomes part of the monthly ritual.

Thresholds should be agreed before the report arrives. If high-potential near misses increase, who reviews them and by when? If critical control checks fail twice in one area, what stops or changes? If observation quality drops, who coaches supervisors? If reporting volume falls in a high-risk operation, who tests for silence?

Without thresholds, leaders negotiate meaning after seeing the number. That makes the interpretation vulnerable to optimism, production pressure, and personal preference. The indicator becomes a talking point rather than a trigger.

A safety KPI owner review is useful here because it assigns a person to definition, interpretation, escalation, and decision quality. Leading indicators need that ownership even more than lagging indicators, because their value depends on timely action.

Gap 4: Signal quality is assumed from volume

The fourth gap is assuming that more reports mean better prevention. More observations, more near misses, more inspections, and more corrective actions can be good news when they contain usable risk information. They can also become noise when people learn to feed the system with easy records.

Volume is seductive because it is simple to compare. One site submitted 800 observations, another submitted 240, and the first one looks more engaged. The comparison may be false if the 800 records repeat low-risk conditions while the 240 records identify weak barriers in serious work.

As Andreza writes in A Ilusao da Conformidade ("The Illusion of Compliance"), good indicators do not guarantee good practices. A leading indicator should therefore be sampled for content quality, not only counted. The record should name the exposure, the control condition, the owner, the action, and the evidence that the action worked.

The danger is not only wasted reporting effort. Poor signal quality teaches leaders to distrust the system, and it teaches workers that safety reporting is about feeding a dashboard. Once that belief settles in, the organization has to rebuild reporting credibility before the metric can lead anything.

Gap 5: Green status is not tested against worker voice

The fifth gap appears when leaders accept green status without asking whether workers and supervisors closest to the task see the same risk picture. A dashboard can be green because reports are late, categories are broad, thresholds are generous, or people have stopped raising concerns.

Worker voice is not a decoration for culture programs. It is a source of risk intelligence. If the people performing the work say a procedure cannot be followed under time pressure, a guard is routinely bypassed, a permit is copied, or a contractor interface is unclear, that evidence should challenge the leading indicator before the meeting accepts green.

This does not mean every concern proves the metric is wrong. It means green status should survive contradiction. A mature review asks what field evidence would disprove the dashboard, then goes looking for it in interviews, observations, control checks, maintenance records, and supervisor debriefs.

The existing article on underreporting distortions shows why silence can make metrics look clean. For leading indicators, the same risk is severe because a silent system removes the early warning that the indicator was built to capture.

Gap 6: The indicator survives after the risk has changed

The sixth gap is indicator inertia. The organization keeps measuring the same activity after the work, workforce, contractor mix, equipment, or risk profile has changed. A leading indicator that was useful during one phase of operation can become irrelevant after a shutdown, expansion, process change, leadership change, or outsourcing decision.

This is common in mature systems because old indicators feel safe. Removing them can look careless, while adding new ones looks responsive. The result is a crowded dashboard where leaders skim familiar charts and miss the fact that the leading indicator set no longer matches the work.

ISO 45001:2018 requires organizations to monitor, measure, analyze, and evaluate occupational health and safety performance. That requirement supports review, not indicator immortality. If the risk has changed, the measurement system should change with it.

An EHS manager should schedule a quarterly retirement review for leading indicators. Keep the indicators that still change decisions, revise the ones whose definitions no longer fit, and remove the ones that only prove activity. A metric that cannot lead a decision should not occupy leadership attention.

How to separate prevention signals from activity theater

The table below gives EHS managers a fast diagnostic. It does not ask whether the organization is busy. It asks whether the activity still creates early, useful, decision-ready risk information.

What EHS managers should change in the next review

The next dashboard review should not start by adding more indicators. It should start by selecting the five leading indicators that receive the most attention and testing each one against six questions. Does it reach serious exposure? Does it verify a control? Does it have a decision threshold? Is signal quality sampled? Can worker voice contradict it? Is it still aligned with the current risk profile?

If one of those answers is weak, the issue is not cosmetic. The indicator may be producing confidence faster than it produces prevention. That is when leaders need to change the metric, the review question, or the field routine behind the number.

For teams connecting indicators with culture, Andreza Araujo's Safety School and books offer a practical base for moving from reporting discipline to leadership behavior. Safety is about coming home, and leading indicators deserve that name only when they help leaders see risk early enough to act.

The useful test is whether the metric gives time back

A leading indicator earns its place when it gives the organization time. Time to repair a control, coach a supervisor, stop a drift, escalate a contractor issue, reopen a weak action, or listen to a concern before harm confirms what the metric should have shown.

The goal is not a larger dashboard. It is a sharper one. Leading indicators should make exposure visible, test the controls that matter, and force decisions while the organization can still choose a better outcome.

Andreza Araujo's resources at andrezaaraujo.com help leaders connect safety culture, metrics, and field behavior. Start with one leading indicator this month and ask whether it gives time back to the people doing the work. If it does not, it is activity dressed as prevention.