Zero-Accident Targets: 5 Distortions in Safety Metrics

Zero-accident targets are corporate safety goals that define success as a period with no recorded injuries or reportable events. They can focus attention on prevention, but they also distort safety metrics when leaders reward a clean number more than honest reporting, weak-signal visibility, and verified risk control.

A zero-accident month can mean excellent prevention. It can also mean that supervisors delayed case classification, workers avoided reporting pain, near misses stayed informal, and executives read silence as control. The danger is not the aspiration to prevent harm. The danger is treating a perfect number as proof that the organization has become safer.

As Andreza Araújo argues in her Portuguese title Muito Além do Zero, glossed as Far Beyond Zero, the most mature safety conversation goes beyond celebrating absence of accidents and asks what the system is learning before harm occurs. Across 25+ years leading EHS in multinational operations, Andreza Araújo has seen that a clean dashboard deserves the same scrutiny as a bad one, because both can hide weak controls when the measurement system rewards appearance.

Why can zero-accident targets become a metric problem?

Zero-accident targets become a metric problem when the target shifts attention from risk reduction to record protection. Once managers are ranked, rewarded, or publicly recognized for a zero line, the organization creates a strong incentive to keep the line clean, even when the work system still contains uncontrolled exposure.

The U.S. Bureau of Labor Statistics reported 5,070 fatal work injuries in the 2024 Census of Fatal Occupational Injuries. That figure matters because fatal exposure can exist inside companies that also show low recordable injury rates. A board that sees only the absence of recordables may miss the conditions that produce severe harm, especially in high-energy work, contractor interfaces, driving, maintenance, and line breaks.

The stronger question for leaders is not whether zero is morally desirable. Of course nobody should be injured. The stronger question is whether the target creates better control of fatal scenarios or simply better protection of the reported rate. That is why the target has to be tested against SIF rate, TRIR, and precursor indicators, not used as a substitute for them.

1. Distortion: underreporting becomes rational

The first distortion appears when workers and supervisors learn that reporting creates trouble while silence protects the local score. A minor strain becomes informal stretching. A first-aid case stays off the system. A near miss is discussed in the locker room but not entered in the database. The dashboard stays clean, although the organization has lost evidence.

This is not a character defect in the workforce. It is a predictable response to a measurement design that treats every reported event as failure. James Reason's work on organizational accidents helps explain why this matters, because latent conditions remain dangerous precisely when information does not travel upward. When reporting is socially expensive, the organization loses the data it needs to see those conditions.

Underreporting is especially likely when zero is tied to bonuses, site rankings, contractor eligibility, or public celebration. In those contexts, the worker who reports pain after a manual handling task may feel responsible for breaking the streak, while the supervisor may feel pressure to solve the matter off system. The company then receives an attractive number and a weaker learning process.

EHS should compare zero months with first-aid logs, clinic visits, restricted-work discussions, overtime peaks, supervisor notes, and anonymous voice channels. If those sources show discomfort while the official dashboard shows nothing, the metric is not clean. It is filtered. The practical audit connects directly with the limits of lagging indicators for EHS decisions.

2. Distortion: leaders confuse absence with control

The second distortion is conceptual. No injury in a month does not prove that critical controls worked. It may only prove that no loss event crossed the reporting threshold during that period. In low-frequency, high-severity risk, luck can look identical to capability until the day a serious event exposes the difference.

Andreza Araújo develops this distinction in Sorte ou Capacidade, glossed as Luck or Capability. The central warning applies directly to executive dashboards: a company can experience a good result without having built the capability that makes the result repeatable. That is why zero should never be read without evidence from control verification, serious-potential near misses, and exposure trends.

Consider a distribution center with one million vehicle movements, no recordable injuries, and repeated pedestrian separation deviations. The zero line may look excellent, but the exposure has not disappeared. If leadership celebrates the rate without asking about vehicle interaction controls, the target has turned into a veil over risk.

The stronger dashboard separates outcome absence from control strength. It asks whether barriers were verified, whether failures repeated, whether supervisors escalated weak signals, and whether serious-potential events received closure with field evidence. That is the line between a lucky month and a controlled month.

3. Distortion: SIF exposure hides behind minor-injury success

The third distortion appears when zero-accident campaigns focus on total injury frequency while serious injury and fatality exposure lives in a different risk layer. A site can reduce slips, first-aid cuts, and low-severity sprains while still carrying uncontrolled energized work, mobile-equipment interaction, dropped-object exposure, and confined-space weakness.

The National Safety Council and Campbell Institute have both emphasized the need to manage serious injury and fatality prevention separately from general injury frequency. The operational reason is simple: the controls that prevent a minor hand cut are not the same controls that prevent a fatal arc flash, a crushed pedestrian, or a fall from height. A single zero line blends those realities until leadership loses severity sensitivity.

For C-level review, the zero target should sit below a serious-risk panel, not above it. The panel should show SIF events, SIF-potential near misses, high-energy task exposure, failed critical-control verification, overdue high-risk actions, and repeat deviations in permit-to-work or isolation. Without that layer, a company can proudly report zero while its fatal scenarios remain untested.

This is where near-miss quality matters more than reporting volume. A serious-potential near miss should challenge a zero celebration, because it reveals the kind of exposure the injury count failed to show.

4. Distortion: supervisors manage the number before the work

The fourth distortion is behavioral. Supervisors under pressure may start managing the reporting boundary before managing the work system. They become skilled at explaining why a case is not recordable, why discomfort is unrelated to work, or why a near miss does not deserve escalation, while the actual exposure remains unchanged.

This does not mean supervisors are dishonest by nature. It means the organization has trained them to treat the metric as a reputation object. When production, bonus, contractor performance, and leadership recognition all depend on the zero line, the supervisor receives a mixed message: prevent harm, but do not let the number move.

During the tenure at PepsiCo South America, where the accident ratio fell 50% in six months under a 180-day plan, Andreza Araújo learned that leadership rhythm changes safety only when it changes field questions. The leader who asks only whether the site is still at zero receives a defensive answer. The leader who asks which control failed this week receives operational evidence.

A practical correction is to remove zero from the first screen of the supervisor dashboard. Start with high-risk work completed, critical controls verified, serious-potential near misses reviewed, and corrective actions closed with field checks. The injury rate can remain visible, but it should not be the first or loudest signal.

5. Distortion: corrective actions become symbolic

The fifth distortion happens after a recordable event breaks the zero streak. Because the event is interpreted as a failure of the target, the organization wants fast restoration of confidence. That often produces symbolic corrective actions: retraining, a new poster, a rewritten procedure, or a toolbox talk that proves activity but not risk reduction.

As Andreza Araújo argues in A Ilusão da Conformidade, glossed as The Illusion of Compliance, formal evidence can hide weak operational reality when leaders accept documentation as proof. In metric systems built around zero, this trap is common because the organization wants to close the visible failure and return to the clean line.

The corrective action should answer a sharper question: what changed in the control system that makes recurrence less likely under production pressure? If the answer is only training, the organization has probably addressed awareness while leaving design, staffing, supervision, maintenance, or planning untouched.

Use action aging and verification as counterweights. A corrective action that remains open for 120 days after a serious-potential event is not only an administrative delay. It is an active risk signal. The method in a corrective action aging dashboard helps leaders see whether the response is changing exposure or only protecting the narrative.

What should replace a zero-only safety dashboard?

A zero-only dashboard should be replaced by a layered dashboard that distinguishes outcomes, exposure, controls, and voice. The goal is not to abandon the moral ambition that nobody gets hurt. The goal is to stop pretending that the ambition can be measured by a single absence.

This structure changes the leadership conversation. Instead of asking whether the site is still at zero, executives ask whether serious exposure is visible, whether controls were verified, whether workers are speaking, and whether actions changed the field. The number then becomes context, not the whole story.

How can executives test whether zero is hiding risk?

Executives can test the target with 5 questions. First, what events became harder to report after zero was introduced? Second, which high-energy controls failed during zero months? Third, which site has the cleanest dashboard but the weakest voice data? Fourth, which corrective actions changed engineering, staffing, planning, or supervision? Fifth, what would the dashboard show if serious-potential near misses carried the same visibility as recordables?

Those questions are uncomfortable because they move the discussion away from celebration and toward evidence. They also protect leaders from the most expensive illusion in safety metrics, which is the belief that a silent system is a safe system. Silence can be maturity, but it can also be fear, fatigue, or filtered reporting.

The better executive habit is to treat zero as an invitation to verify. When a site reports a perfect month, leaders should ask for the control evidence behind the result. If the evidence is strong, celebrate it. If the evidence is thin, the zero line becomes a prompt for investigation rather than a trophy.

Zero should remain an ambition, not the metric architecture

No serious safety leader wants injuries. Zero can remain a moral ambition because every person should return home healthy. It should not become the metric architecture, because architecture tells people what information has value, what can be ignored, and what must be protected.

The 5 distortions are connected by one pattern: when zero becomes the headline measure, people learn to defend the line. Mature safety leadership asks for something harder. It asks whether risk is visible before harm, whether controls work under pressure, whether supervisors can tell the truth, and whether executives reward evidence even when the evidence is uncomfortable.

That is the practical meaning of going far beyond zero. The organization does not lower its ambition. It raises the quality of proof.