Leadership behaviors that matter

Three weeks after the spiral conveyor injury, Rachel Okoye asked every manager and supervisor at Harrowgate Foods Inc. to write down the three things they had done in the past month that showed their crews safety mattered. Most lists were short and similar: 'attended the safety meeting', 'signed off the monthly inspection', 'reminded the crew about PPE'. Then she asked the same question of 40 front-line workers across the three sites, about their own supervisor. The most common answer was 'I don't know' and the second most common was 'he tells us to be safe at the start of the shift'.

The exercise made a simple point. Leaders often believe they are showing commitment to safety because they attend the right meetings and say the right words. Workers judge commitment by different evidence: what leaders do when it costs them something. This lesson looks at the leadership behaviors that research and experience link to stronger safety cultures, and at why consistency between words and actions matters more than any single gesture.

Why leaders shape culture

Schein argued that the most powerful way leaders embed culture is through what they systematically pay attention to, measure and control; how they react to critical incidents and crises; how they allocate resources; what they role-model and coach; how they allocate rewards and status; and how they recruit, promote and remove people. Speeches and posters come much lower down his list. People are constantly reading these signals, because they need to know what is really expected to keep their jobs and get ahead.

This is true at every level. Senior leaders set priorities, budgets and targets. Middle managers translate them into schedules and staffing. Front-line supervisors decide, shift by shift, whether a crew stops or keeps going. Studies of safety climate consistently find that the immediate supervisor has the strongest influence on how workers behave, because the supervisor is the person whose reactions they see every day.

Leadership styles and safety

Researchers have studied which leadership styles are linked to better safety outcomes. Two broad patterns stand out.

• Transactional leadership works through clear expectations, monitoring and consequences: set the rule, check compliance, reward or correct. It is necessary, especially for critical rules such as lockout, and it is how most safety systems are built.
• Transformational leadership works through vision, example, individual attention and intellectual challenge: explaining why safety matters, showing it personally, caring about individual workers and inviting them to question how work is done. Studies across industries have linked transformational safety leadership to higher safety participation, better climate scores and fewer injuries.

Good safety leaders use both. They are clear and firm about the few rules that protect people from death and serious injury, and they engage, explain and listen about everything else. A leader who only enforces gets compliance while watched. A leader who only inspires, without holding anyone to account, gets warm feelings and drift.

The behaviors that make the biggest difference

Elena Novak pulled together the leadership behaviors most often linked to strong safety cultures in research and in the experience of companies that had improved. She grouped them into six areas and used them to build a leadership expectation for Harrowgate.

Be present where work is done. Leaders who spend time on the floor, at the docks and on night shift see how work really happens and show that they care enough to come and look. This is sometimes called visible felt leadership: not just being seen, but being felt as someone who genuinely cares. Lesson 4 covers how to make these visits useful.

Ask and listen. Good safety leaders ask open questions such as 'What makes this job hard?' and 'What would you change if you could?' and then listen without defending. Workers know the hazards and workarounds better than anyone. Leaders who do most of the talking learn little.

Act on what you hear, and close the loop. Nothing destroys trust faster than asking for concerns and then doing nothing. Not every concern can be fixed immediately, but every concern deserves an answer: what will be done, by when, or why not. Closing the loop is how workers learn that speaking up is worth it.

Protect safety when it costs. The moments that define a leader are when safety costs time, money or output. Stopping a line, delaying a shipment, buying the better guard or adding a second person to a task sends a message that no poster can match. So does backing up a supervisor or worker who stopped a job, even when it turns out the job was safe.

Recognize good safety work. Noticing and thanking people for reporting hazards, stopping unsafe work, improving a procedure or helping a new worker builds the behaviors you want. Lesson 7 covers how to recognize safety without creating incentives to hide injuries.

Be consistent and fair. Workers watch whether the same rules apply to everyone, including supervisors, high performers and managers, and whether leaders react the same way to bad news on a busy day and a quiet one. Inconsistency teaches people that safety is negotiable.

Say-do gaps

The most damaging leadership signal is a gap between what leaders say and what they do. Every gap teaches people that the real rules are different from the stated ones. At Harrowgate Foods Inc., the investigation and the follow-up interviews found several:

None of these gaps was intended. Each made sense to the person involved at the time: the shift manager was worried about the oven line, the managers were only on the floor for a few minutes, the distribution center manager thought retraining was supportive. But workers do not see intentions. They see actions, and they draw conclusions.

Closing a say-do gap usually needs two things: changing the behavior, and saying out loud that it has changed. When Marcus Bell moved safety to the top of the morning meeting, he told the supervisors why: 'I have been sending the wrong message for two years by asking about restarts first.' Owning the gap publicly made the change credible.

Commitment shows in decisions about resources

Leaders also show what they value through budgets and staffing. After the injury, maintenance proposed moving the spiral conveyor disconnect to the floor beside the machine and adding a second sanitation worker on nights. Together these would cost about the same as one week of overtime. They had been proposed twice before and deferred both times. Rachel approved them within a day and told the night crew directly. More than any speech, that decision told the night shift that something had changed.

Leadership commitment is not just good practice; it also connects to legal exposure. Under the federal Occupational Safety and Health Act (OSH Act), the employer is responsible for compliance, and OSHA treats a supervisor's knowledge of a hazard as the employer's knowledge. Willful and repeat violations carry much higher penalties, and a willful violation that causes a death can be referred for criminal prosecution. OSHA's recommended practices for safety and health programs put management leadership first among the core elements, describing it as top management demonstrating commitment, setting goals, allocating resources and setting an example.

The supervisor's special role

Front-line supervisors face the sharpest goal conflicts. They are measured on output, quality, attendance and cost, and they are told safety comes first. Most of them were promoted because they were good at the job, not because they were trained to lead. Supporting supervisors is one of the most effective things senior leaders can do for safety culture. At Harrowgate Foods Inc. that meant:

• making it explicit, in writing and in person, that a supervisor who stops work for safety will be backed even if production suffers;
• giving supervisors time in the schedule for pre-shift talks, walk-arounds and coaching;
• training them in how to have safety conversations, respond to reports and apply fair accountability;
• removing measures that rewarded supervisors for low reported injury counts;
• asking supervisors regularly what gets in the way of working safely, and acting on the answers.

Tracking leadership behaviors

Rachel wanted to know whether leaders were actually changing what they did. She asked each of the 22 managers and supervisors to log four behaviors each week: floor visits including at least one on a back shift, safety conversations held, concerns closed with an answer to the person who raised them, and recognitions given. After eight weeks, Elena summarized the logs.

# Eight weeks of leader activity logs (totals across 22 leaders).
weeks = [1, 2, 3, 4, 5, 6, 7, 8]
visits = [18, 25, 31, 36, 40, 41, 44, 46]
backshift = [2, 4, 7, 9, 12, 13, 15, 16]
conversations = [35, 52, 70, 81, 90, 94, 99, 103]
raised = [14, 20, 24, 27, 26, 25, 24, 22]
closed = [3, 8, 13, 18, 21, 22, 22, 21]
for w, v, b, c, r, k in zip(weeks, visits, backshift, conversations, raised, closed):
    print(f"Week {w}: visits {v:2d} ({b:2d} back shift), conversations {c:3d}, "
          f"concerns closed {k:2d}/{r:2d} = {100*k/r:3.0f}%")
print(f"Back-shift share rose from {100*backshift[0]/visits[0]:.0f}% to {100*backshift[-1]/visits[-1]:.0f}%")

Week 1: visits 18 ( 2 back shift), conversations  35, concerns closed  3/14 =  21%
Week 2: visits 25 ( 4 back shift), conversations  52, concerns closed  8/20 =  40%
Week 3: visits 31 ( 7 back shift), conversations  70, concerns closed 13/24 =  54%
Week 4: visits 36 ( 9 back shift), conversations  81, concerns closed 18/27 =  67%
Week 5: visits 40 (12 back shift), conversations  90, concerns closed 21/26 =  81%
Week 6: visits 41 (13 back shift), conversations  94, concerns closed 22/25 =  88%
Week 7: visits 44 (15 back shift), conversations  99, concerns closed 22/24 =  92%
Week 8: visits 46 (16 back shift), conversations 103, concerns closed 21/22 =  95%
Back-shift share rose from 11% to 35%

The logs showed real progress: floor visits more than doubled, back-shift visits rose from about one in nine to about one in three, and the share of concerns closed with an answer rose from about one in five to almost all of them. They also showed something Elena flagged: the number of concerns raised peaked in week 4 and then started to fall. That could mean fewer problems, or it could mean people had raised the obvious issues and were waiting to see whether the hard ones got fixed. Behavior logs are a useful prompt, but they measure activity. Lesson 11 covers how to check whether that activity is changing what workers experience.