How Dr. Thomas Krause Thinks About Leadership Quality Before Safety Metrics Move

Episode 11 of the Headline Podcast, published on December 3, 2025, featured Dr. Thomas Krause, CEO of the Krause Bell Group. His central argument was that safety performance changes when leaders improve the quality of decisions and relationships around the work, rather than simply declaring that safety matters.

That argument shifts the leadership question from “How visible is the safety message?” to “What does leadership make easier, harder, faster, or safer every day?”

Why leadership quality comes before safety metrics

Leadership quality comes before safety metrics because the quality of leadership shapes whether people can identify exposure, act on weak signals, and learn from failure. A metric can show that performance moved, but it cannot explain which decisions created that movement. Leaders who want a useful safety result must examine how priorities are set, how concerns travel, and how supervisors respond when production pressure conflicts with protection.

Dr. Krause stated on the episode, “We were surprised to learn the strongest predictor of success was the quality of leadership given to the initiative.” The point is not that leaders personally control every outcome. It is that they establish the conditions in which controls are funded, procedures are practical, and difficult information can reach a decision-maker before the exposure becomes normal.

That distinction matters because organizations often review the final number while ignoring the management choices that preceded it. A lower injury rate may reflect better control, lower exposure, underreporting, or a change in work mix. A higher reporting rate may signal failure, or it may show that people trust the response enough to make weak signals visible.

Andreza Araujo’s work across more than 250 cultural transformation projects supports this leadership reading. The useful question is not whether executives have communicated a value. It is whether the value is present in the decisions that shape staffing, maintenance, supervision, contractor control, and recovery after a disruption.

What leaders make visible on the floor

Leaders make safety credible when they spend time where decisions are executed, ask questions that reveal conditions, and return with a visible response. A floor visit is not a performance of concern. It is a test of whether leadership can see the gap between the approved system and the work that people must actually complete.

Dr. Krause connected leadership with walking the floor and asking the right questions. That practice is more demanding than a tour in which every answer is accepted and every observation becomes a generic action item. A useful visit asks what makes the safe method difficult, which decision is delayed, what people do when the planned control is unavailable, and who has authority to change the condition.

The answer often sits outside the safety department. A supervisor may know that a control is weak but lack budget authority. A maintenance planner may understand the exposure but have no protected window for correction. A contractor may recognize the hazard but believe that stopping the task will threaten the relationship. Leadership quality appears in how quickly those constraints become a management decision.

Use voice climate signals to check whether bad news can travel, then compare the findings with the decisions leaders made after the last visit. If the same concern appears three times without a changed owner, resource, or control, the visit created visibility without governance.

The 60 percent trust trap

A favorable trust score can hide a serious leadership weakness when the remaining minority is exposed to different conditions or cannot speak safely. Dr. Krause used a 60 percent trust example to show why averages can flatter leaders. The question is not whether the score looks acceptable, but whose experience the average hides and what consequence follows.

On the podcast, Dr. Krause said, “You could have 60% of people say they trust their supervisor and still be in the 90th percentile, so the leader pats himself on the back while 40% of his people don’t trust their boss.” The number is not a universal threshold. It is a warning against converting a distribution of experiences into one reassuring headline.

A leader should separate the result by shift, tenure, contractor status, location, language, and proximity to the highest-consequence work. The 40 percent may not be evenly distributed. It may be concentrated among night-shift workers, new starters, or people who must challenge a production decision without direct access to senior leadership.

Pair a trust measure with observable evidence. Review how many concerns received a named response, how long decisions took, whether controls changed, and whether reporters were told what happened. The pattern matters more than a single score because trust is built through repeated consequences, not through the survey instrument alone.

How the leadership view differs from a metric-first system

A metric-first system asks whether the result is within target. A leadership-quality system asks which decisions produced the result, who was affected, and whether the conditions are likely to hold under pressure. Both can use numbers, but the second system treats numbers as evidence for inquiry rather than as a substitute for judgment.

The difference is not an argument against measurement. It is an argument for better questions around the measurement. A dashboard can show that corrective actions closed in 30 days, yet a field review may reveal that the same workaround remains necessary because the design decision was never funded.

Leaders can compare their dashboard with a field review and capital gate to see whether financial and operational decisions tell the same story as the safety data. When they diverge, the divergence deserves attention instead of a cosmetic adjustment to the dashboard.

Why the incident story often points backward

An incident investigation becomes a leadership test when it asks why the work was difficult to perform as intended. The immediate action may involve a person, but the management explanation often reaches backward through planning, design, staffing, incentives, and earlier decisions that made the unsafe path more available.

Dr. Krause argued that incident analysis can make the employee appear to be the whole problem when leaders ignore system factors set by decisions made a year, or five years, ago. That five-year reference matters because leadership creates conditions that outlast the meeting in which they were approved.

This does not remove personal responsibility. It places responsibility at the level where the condition can be changed. If a procedure requires a control that the equipment cannot provide, the corrective action belongs with design and capital governance. If a supervisor is expected to stop work but is measured only on schedule recovery, the leadership system is sending two conflicting instructions.

Use post-incident meeting questions to test whether the organization is learning about the work or only documenting a violation. Ask which decisions made the exposure persistent, what earlier signal was available, and which owner can prevent recurrence without relying on perfect human behavior.

What executives should listen for in the next review

Executives should listen for the language of decision quality. Strong reviews identify an exposure, a decision owner, a constraint, a control condition, and a verification date. Weak reviews rely on broad assurances, repeat the target, or describe culture as if it were separate from the work choices that leaders control.

Listen for whether the discussion distinguishes a declared priority from a protected priority. A declared priority appears in speeches, posters, and scorecards. A protected priority appears when a leader accepts delay, funds a redesign, changes a staffing assumption, or asks a contractor to stop because the control is not credible.

Listen for whether people can describe the last time a concern changed a decision. This is more informative than asking whether the organization encourages speaking up, because encouragement is an intention while a changed decision is evidence. If nobody can name an example, the communication strategy may be ahead of the response system.

Listen for the time horizon. A leader who asks only what happened this month may miss conditions created one, three, or five years earlier. A leader who asks what will be different next month without naming an owner is also leaving the work unfinished.

Recommendation

Use the next safety review to examine leadership quality through three linked questions. What decision created or tolerated the exposure? Which people experience the system differently from the average? What visible change will prove that leadership acted? These questions turn the episode’s argument into a practical review method.

Start with one material risk and one recent decision. Ask the executive sponsor to name the owner, the constraint, the authority required, and the date for field verification. Then speak with the people closest to the work, including the group whose trust may be hidden by the average. Record what they say without converting every concern into a promise that cannot be delivered.

Review the result after 30 days. If the exposure has changed, document the management choice that made the change possible. If it has not changed, escalate the constraint rather than relabeling the action as complete. A leadership review earns credibility when it leaves behind a different decision, not merely a better explanation.

Andreza Araujo’s book Safety Culture: From Theory to Practice makes the same practical connection between culture and observable management behavior. For a broader view of how the Headline Podcast approaches leadership and safety, read the Headline Podcast editorial library, and explore Andreza Araujo’s books for the frameworks that support this work.