What we all would like to know: Does Leadership Influence Business KPIs?

Dutch Railways (NS) has about 1,000 managers. What makes these managers effective? To answer that question Hanneke Beelen (Leadership) and Jean-Paul Lucassen (People Analytics) of NS worked together with the University of Amsterdam to study the effectivity of leaders in operations. Read on to learn about their findings.

The impact of good leadership

Leadership is a contemporary topic, especially the role of the leader. What do we expect of our managers? What is their objective and how can NS facilitate them to achieve that objective?

NS was particularly curious about whether team managers had an influence on KPIs like customer satisfaction, performance (like driving trains on time), absenteeism, and employee behavior (which is measured in surveys).

We also wanted to know if there is any difference between two styles of leadership: empowering leadership and task-oriented leadership.

• Empowering leadership means that the leader gives autonomy and motivates his employees by inspiring them.
• Task-oriented leadership means that the leader focuses on results.

We specifically chose our team managers (353 FTE) as our population of study because they have a responsibility in the primary process of NS and therefore directly influence the business KPIs.  

Based on these questions and literature study, the Amsterdam People Analytics Center (APAC) of the University of Amsterdam formulated hypotheses that both styles of leadership have a positive effect on the KPIs, except for absenteeism where you would expect a decline.

Combining HR data and Business Data

For NS, this was the first time we conducted research by combining HR data and business data. The collected data were processed to make sure it was not possible to trace back to an individual level. We did this in accordance with the guidelines of our Data Usage Board, an internal organ that helps NS to be GDPR compliant.

We used three years of historical data and used several attributes regarding our team managers: age (in cohorts), years of service, job title, number of direct reports, years at position, and performance score. We also added how the managers scored their employees on performance, meaning we added the distribution of performance scores in the team.

The next dataset we added was the results of our engagement survey on the team level. This survey, for example, measures how employees think about being proud of working for NS and whether they think they are managed in a pleasant way. On the team level, we added the delta in absenteeism, meaning we added the incline or decline in absenteeism in periods of six months. With three years of data, we had six absenteeism delta’s per team. 

To combine the HR data with business data we then added data on the business KPIs on the team level. Finally, we added the results of a survey by APAC about the level of empowering leadership that was being experienced.

Successful behavior

Organizational success depends largely on the behavior of employees. Nonetheless, there are more factors that influence this success. NS is developing a new performance model that gives insights into what behavior is successful in different positions. The model contains three elements: the results for which one is responsible, the context in which someone is working and the capacities an employee brings to work. For team managers, successful behavior was also specified. This study would act as a validation of this performance model.

Results

• A strong correlation was found between leadership style and absenteeism. The absenteeism delta of a team can be explained for an important part by the style of leadership of a manager;
• In general team managers scoring high on empowering leadership have a stronger decline in absenteeism. Task-oriented leadership, on the other hand, leads to an increase of absenteeism;
• How (positive and negative) and in what amount leadership style influences the absenteeism delta was different for teams with specific functions (drivers, conductors, etc.);
• The effectivity of leadership behavior fluctuates per function, but also context (teams with high and low absenteeism).

For the other KPIs, we found no correlations. A possible reason for not finding these correlations could be data noise. We used different sources filled in by different people like employees, managers, and customers. It could also be because we used aggregated data and analyzed it on a team level.

Results on Empowering leadership

The study shows that empowering leadership is appreciated widely, both by employees and managers. When team managers score high on empowering leadership, we see that:

• Performance scores of managers are higher;
• Employees have a more positive perception regarding their team manager;
• Well-being (a combination of engagement questions) of employees in the team is higher;

We also found that successful behavior is a good way to manage empowering leadership.

What’s next

The results gave us input for a follow-up study and might eventually lead to new policies and interventions:

• NS has started a follow-up study on absenteeism. Further research focusses on empowering leadership and instead of using absenteeism delta’s we want to use actual absenteeism data;
• Employee engagement is an important data source. After analyzing the items and answers, we found that the current survey was not usable enough for this study. A new tender regarding employee engagement is running and we will design new questions regarding leadership for future surveys;
• Successful behavior: NS is using a pilot to study how success factors can be used in employee conversations to manage effective behavior more focused. We think the outcomes of the pilot will help us to improve both the management and performance models, and give us a clear view of what skills or competencies are important in our work.