The Importance of Quality Data for Continuous Improvement in Safety Management Systems
In the world of occupational safety, the phrase “you can’t improve what you don’t measure” holds especially true. But not all measurements are created equal. For a Safety Management System (SMS) to mature and evolve, it must be built upon quality data—data that not only reflects the current state of the organization but also points to where risk lives and where interventions can make a measurable difference. Unfortunately, many organizations still rely heavily on lagging indicators, or worse, adopt misleading pseudo-leading indicators that offer the illusion of progress while concealing critical system weaknesses.
Lagging indicators—like recordable injury rates, days away from work, or lost-time incident frequency—only tell us what happened after something went wrong. While these metrics are easy to quantify and compare, they are inherently reactive. They do nothing to uncover the systemic issues that precede incidents, and they only provide feedback after the damage has been done. In many cases, an overemphasis on lagging data can even encourage underreporting and a false sense of security.
Recognizing the limitations of lagging indicators, safety professionals have turned to leading indicators. However, many so-called leading indicators are merely activity measures. Take for example the common metric “percent of workforce that completed safety training.” While it appears forward-looking, it does not measure whether the training was effective, retained, or applied. A team could have 100% training completion and still experience serious incidents if the training was poorly designed, irrelevant, or delivered in a disengaging way.
Being active is not the same as being effective.
A true leading indicator should do more than measure output—it should provide insight into how well the system is functioning and where improvements can be made. High-quality leading indicators share several critical qualities:
Diagnostic Value – A good leading indicator reveals weaknesses or inefficiencies in safety processes before an incident occurs. It should help identify specific failure points in training, hazard controls, procedures, or communication.
Predictive Power – It should have a demonstrated relationship to safety outcomes, showing that changes in the indicator reliably precede changes in risk or incident frequency.
Actionability – A useful metric provides enough clarity and specificity that organizations can act on it directly. For example, tracking the percentage of corrective actions closed on time or the average time to resolve near-miss reports provides insight and a path forward.
Proactive Risk Reduction – It should reflect the organization’s efforts to reduce risk before harm occurs, such as identifying hazards, closing safety work orders, or auditing high-risk tasks.
Engagement-Focused – Good indicators measure the quality of participation, not just quantity. This might include tracking involvement in safety observations, peer-to-peer feedback, or leadership participation in hazard walkthroughs.
Process-Oriented – Instead of focusing only on outcomes, a strong indicator assesses how well safety processes are implemented and followed, such as adherence to permit-to-work procedures or pre-task risk assessments.
The future of safety performance lies in the strategic use of high-quality data. By moving away from vanity metrics and toward meaningful, actionable indicators, organizations can achieve true continuous improvement in their safety management systems—not just activity for activity’s sake, but real progress built on insight, engagement, and accountability.
