Why Safety Culture Metrics Can Mislead Leadership
A falling incident rate can mean the site is safer. It can also mean people have stopped reporting. The board dashboard should be able to tell which.
What the incident rate measures
Recordable incident rates are lagging indicators that count injuries after they have happened. Boards rely on them because they are standardized and comparable across sites and years.
They are also exposed to a quiet distortion. Suppose frontline workers believe a near-miss report will count against their shift, so they report less and minor injuries are handled informally. The rate improves, and nothing about the hazards has changed.
From the boardroom, that looks identical to real improvement.
Choosing leading indicators
A leading indicator moves before harm occurs, and the useful ones have two properties. They respond to conditions on the floor, and they are hard to improve by staying silent.
Exhibit 1: Common safety indicators, what each tells a board, and how each can mislead.
| Indicator | Type | What it shows | How it misleads |
|---|---|---|---|
| Recordable incident rate | Lagging | Harm that occurred | Falls when reporting is discouraged |
| Near-miss reports filed | Leading | Hazards noticed and raised | A drop can mean fear, not safety |
| Time to close corrective actions | Leading | Whether raised hazards get fixed | Actions closed on paper only |
| Safety-critical maintenance overdue | Leading | Condition of key barriers | Says nothing about behavior |
| Supervisor response, from frontline interviews | Leading | Whether it is safe to speak up | Needs independent collection |
No single figure is enough, but a dashboard that pairs a lagging measure with two or three leading ones lets directors see when the numbers disagree.
Five questions for the board
- Has near-miss reporting moved in the same direction as the incident rate, or the opposite one?
- When did someone independent of line management last interview frontline crews?
- Does any incentive reward a site or shift for reporting nothing?
- How many reported hazards from last quarter are still open?
- What happened to the last person who reported a serious near-miss?
Finding out what is really happening
Underreporting can be detected, though seldom by the compliance audit, because people questioned in an audit describe the system as designed. Anonymous surveys, interviews and watching how supervisors react to a raised concern give a truer picture.
In one composite engagement, a manufacturer with incident rates inside industry norms found crews were withholding reports. After supervisors were coached on how to receive them, near-miss reporting more than doubled within six months. Read the safety culture case study.
This article takes the board's view. For the signs a supervisor can spot on the floor, and how to rebuild trust in reporting, see when near-miss reports stop.
Changes to make to the dashboard
- Add near-miss report volume to the board pack and treat a sharp fall as a question to investigate.
- Pair each lagging indicator on the dashboard with one leading indicator.
- Commission a culture assessment that is separate from the compliance audit.
- Remove any bonus or league table that rewards a shift for having zero reports.
Test what your safety metrics are telling you
We can review your board safety dashboard and run a culture assessment that shows whether reporting reflects conditions on the floor.
What ISO 45001 Auditors Look For
Certification audits rarely fail a system for missing documents. They fail it when the documents and the site do not match.
Organizational Resilience Amid Regional Instability
Continuity plans are written for a fire or an outage on your own site. Regional disruption arrives from outside and breaks the assumptions those plans rest on.
Why Process Safety Management Fails Quietly
A HAZOP study is accurate on the day it is signed. Each modification afterwards is a small bet that nothing relevant has changed.