The Senior Leader's Distance Problem
Why EHS metrics stop moving — and what executives actually own.
There is a pattern I have seen in more than thirty years of senior safety work that I now consider the defining problem of executive safety leadership in this era. It goes like this.
An organization invests heavily in safety. It builds out a corporate function, rolls out a management system, hires a credentialed safety leader, and gives that leader visibility and budget. The injury rate falls. Sometimes meaningfully, sometimes dramatically. And then, three to seven years in, the rate stops falling. The next set of investments produces less. The next initiative lands flat. The senior leadership team gets frustrated, presses the safety function for new ideas, approves the new ideas, and watches them produce roughly the same plateau, just with different language attached.
I want to make a claim that may be uncomfortable for the executives who read this. The plateau is not a safety function problem. It is a senior leadership problem. And the data backs this up more clearly than most senior teams realize.
Consider what the lagging indicators actually show. In 2024, the U.S. Bureau of Labor Statistics recorded 5,070 fatal work injuries, a 4 percent decline from 2023, and a fatality rate of 3.3 per 100,000 full-time equivalent workers, down from 3.5. Real progress. At the same time, ISN's 2025 multi-framework analysis of more than 178,000 OSHA recordable incidents from 2017 to 2024 found that while overall recordable rates have steadily declined, serious injuries and fatalities have not followed the same trajectory. As ISN's senior vice president Joe Schloesser put it, "Leading organizations with mature safety cultures recognize historical recordable injury rates, such as TRIR, to be less reliable indicators for reducing serious incidents and fatalities." The Campbell Institute and the Edison Electric Institute have reached similar conclusions through independent work — high-frequency, low-severity incidents and SIF events correlate weakly. An organization can drive its TRIR to a record low and still experience a fatality, because the two metrics measure different things produced by different decisions.
Those decisions live outside the safety function. They are made by people whose primary accountability is operational. And the cumulative pattern of those decisions, more than any safety program, determines whether a culture matures past the dependent stage or stalls.
Let me make this concrete. A plant manager who consistently pushes maintenance windows to protect production output is making a safety decision, whether or not it appears on any safety scorecard. A VP who tolerates a culture in which raising a concern produces career consequences is making a safety decision. A CFO who treats safety capital requests as discretionary spending in tight quarters is making a safety decision. A CEO who praises the team that hit the quarterly number without asking how they hit it is making a safety decision. None of these decisions are made by the safety leader. All of them shape what is possible for the safety leader to achieve.
This is the distance problem. The further a leader is from the work, the more abstract safety becomes — and the easier it gets to make a decision whose safety implications are real but invisible to the person making it. The plant manager sees a production number. The CFO sees a capital request. The CEO sees a quarterly result. The worker sees a guard they have to bypass, a maintenance issue that did not get fixed, a near-miss that did not get reported because of the pressure they felt. The chain from the executive decision to the worker's reality is long enough that almost no one connects the dots in real time. The next incident, when it comes, will look like an operational anomaly. It is almost never an anomaly. It is the cumulative residue of decisions made at distance.
What does it look like to own this honestly as a senior leader? I would offer four practices that, in my experience, separate the executives whose organizations break through the plateau from those whose organizations do not.
The first is treating safety as a leadership system, not a function. The safety leader runs a function. The senior team runs the system in which that function operates. The function can deliver excellent work and still fail to produce results if the surrounding system rewards behaviors that contradict the safety message. Senior leaders who understand this stop asking the safety function for new programs and start asking themselves what the system is actually rewarding. The honest answer is often unflattering, which is exactly why most senior teams skip the question. Sidney Dekker has made this point repeatedly in his work on safety differently — leaders have to be willing to see what the organization is actually producing, not what they wish it were producing.
The second is being specific about what executives personally own. "Safety is everyone's responsibility" is true, and it is also a way of letting executives off the hook for the specific decisions only they can make. A CEO does not own PPE compliance. A CEO owns whether the organization's compensation structure rewards or punishes leaders who slow production to address a real hazard. That is a different kind of accountability, and most executive teams have not been asked to articulate it.
The third is closing the distance physically. The leader who shows up on a production floor twice a quarter, unannounced, without an agenda, and stays long enough to actually talk to workers, makes different decisions in the boardroom than the leader who has not been to a site in eighteen months. This is not symbolism — it is information acquisition. The data you carry in your head when you make a capital allocation decision is qualitatively different if you have stood next to the person who will use the equipment. Most senior leaders underestimate this and most never test it.
The fourth is changing what gets asked about. In most senior leadership reviews, safety appears as a slide near the front and a couple of numbers. Recordable rate. Days away. Maybe a leading indicator. The senior team nods, asks one or two questions, and moves on. Now imagine a different review, in which the senior team asks the operating leaders, every quarter, what specific decisions they made in the last ninety days where safety considerations conflicted with operational pressure, and how they resolved them. That single shift in inquiry would change the behavior of operating leaders more than any program rollout. It tells the system that the executives are watching the part that actually matters.
I want to close on a number that should sit in front of every senior leadership team in the country. In 2024, 5,070 workers died on the job in the United States. A worker every 104 minutes. The fatality rate's slight decline is real progress, but the absolute number remains a moral problem rather than a statistical one. The further fall in that number, year over year, will not come from the safety function alone. It will come from senior leadership teams that close the distance and own what is theirs to own.
The safety leader will tell you what the data says. The worker will tell you what the day is like. The decision about which one of those informs your next move is yours alone. That decision is the actual job.
Sources & Further Reading
• U.S. Bureau of Labor Statistics. (2026). Census of Fatal Occupational Injuries 2024.
• ISN. (2025). Serious Injury & Fatality Insights Multi-Framework Analysis 2017–2024.
• Campbell Institute. SIF Prevention research series.
• Edison Electric Institute. (2024). Approved Changes to SIF Criteria.
• Dekker, S. (2014). Safety Differently: Human Factors for a New Era. CRC Press.
