Contrasting insights: digital transformation through the lens of higher versus middle management

The public sector is undergoing rapid changes, as is the character and significance of public leadership. These alterations necessitate engaging with new ideas that could help governments address current institutional and societal challenges (Virtanen and Tammeaid, 2020). In light of these substantial transformations, scholars now widely concur that leadership is vital for the efficiency and accountability of public organizations and therefore merits greater focus (Vogel and Masal, 2015). Among these transformations, digitalization has emerged as one of the most profound forces reshaping public sector leadership, fundamentally changing how leaders must think about and execute their roles (Fernandez-Vidal et al., 2022).

Digital transformation represents an immense shift to how public organizations operate and deliver services (Shen et al., 2023). It includes two interrelated yet distinct leadership challenges: (1) leading in a digitally transformed environment, involving remote work, virtual collaboration, and digitally enabled workflows; and (2) leading digital transformation efforts, which focus on systemic change and strategic implementation of technology (Mergel et al., 2019; Meijer, 2015). These dual challenges require distinct leadership competencies to navigate both operational and strategic dimensions (Khanom, 2023; Sewpersadh, 2023).

Successfully addressing these dual challenges requires the involvement of both middle and higher managers, though their perspectives on organization and implementation often differ significantly (Bäck et al., 2020; Kaiser et al., 2011; Krausert, 2014). Higher managers concentrate on strategic and systemic concerns, such as forming external partnerships, engaging in discussions with regulatory bodies, and establishing support frameworks. Middle managers, conversely, are concerned with practical execution at the employee level, including motivating and engaging staff members (Luthans et al., 1985).

The complexity of integrating digital technologies into public services stems from its dual nature: it requires both technological adoption and cultural shifts within the organization (Newman et al., 2022; Tabrizi et al., 2019). In this context, higher managers' primary responsibility is conveying the urgency of digital transformation, illustrating its connection to other enhancements (Bresciani et al., 2021), and securing requisite resources and support to ensure sustainable transformation across the organization. They also serve as crucial bridges between organizational operations and broader systemic levels, including regional and national frameworks (Klein, 2020; Razzak et al., 2022). Middle-level managers, with their deep understanding of staff needs, oversee daily progress toward implementing digital transformation (Henderikx and Stoffers, 2022; Kaiser et al., 2011). They significantly influence the organization by modeling exemplary behavior and leadership to promote change (Hermkens, 2021). Their dual role as agents of change and implementors of digital strategies places them in a unique position to influence the success of digital initiatives (Kaiser et al., 2011; Luthans et al., 1985).

This pivotal position of middle managers demands comprehensive leadership skills, including technical proficiency, strategic thinking, and the ability to foster collaboration and innovation. In the digital era, crucial leadership skills and behaviors include adaptability, digital literacy, results-oriented management (Klein, 2020; Porfírio et al., 2021), cooperation, self-management, and the ability to navigate change and handle ambivalence (Henderikx and Stoffers, 2023). However, gaps between the skills needed and those that middle and higher managers perceive as necessary lead to disconnects in leadership development strategies (Kaiser et al., 2011; Spagnoli et al., 2021), potentially undermining digital transformation efforts.

These differing perspectives manifest in organizational politics and competing interests between middle and higher managers (Obidike, 2019; Pfeffer, 1993), affecting resource allocation, performance measurement, and ultimately, transformation success. Competition for resources and disagreements over performance indicators create misalignments and tensions, highlighting the need for strategies that foster alignment and collaboration (Sun and Anderson, 2012). Successful digital transformations in large public organizations thus require careful consideration of leadership competencies, organizational dynamics, and alignment among management levels (Fernandez-Vidal et al., 2022; Schiuma et al., 2021). This necessity underscores the importance of developing a unified approach to digital transformation that bridges disparities in perceived leadership skills and behaviors. Given the broad range of required skills and behaviors, there is a clear need to cluster them to enhance comprehensive leadership development strategies. By fostering cultures of continuous learning, public organizations can better navigate the complexities of the digital era (Simmons and Yawson, 2022), ultimately improving service delivery and operational efficiency.

Research on this topic is undeveloped; therefore we designed a study that explores perceived important leadership skills and behaviors given the continued digitalization in organizations. We drew from the experiences of practitioners in the largest public organization in the Netherlands, linking findings with leadership and digital transformation literature. We therefore address the following research questions:

Considering a focus on practice, we designed a study that avoids imposing extant theories of leadership, since such theories’ goals, contexts, and actors differ, making applicability to digital innovation uncertain. Following Blumer’s (1954) recommendations for inductive empirical research, we used sensitizing concepts from relevant literature to interpret results that served as points of departure during analysis. This study adds to earlier research on middle managers by Henderikx and Stoffers (2023). We expanded the data set by including statements from higher managers and re-analyzed the data to investigate new cluster structures and whether disparities exist between middle and higher managers.

Group concept mapping (GCM) is a mixed-methods approach that reveals insights into a group’s comprehension of multifaceted phenomena and discovers novel meanings (Kane and Trochim, 2007). Rosas and Kane (2012), van Bon-Martens et al. (2014), Rosas and Ridings (2017), and Trochim and McLinden (2017) discuss the method extensively, and we therefore only provide a summary of it here. Trochim (1989) indicates that GCM uses participant data to create visualizations that represent a target group’s ideas and opinions. The process follows five steps—preparation, generating statements, structuring statements, data analysis, and data interpretation (Figure 1; adapted from Hagell et al., 2016).

Step 1 comprises participant selection and determining the focus of the study. The researcher creates a focus prompt that is, in fact, a sentence that needs to be completed to guide a brainstorming phase and establish the criteria for rating statements. The rating criteria depend on the study’s purpose. During Step 2, the brainstorming phase, participants generate statements by completing the formulated focus prompt multiple times. The goal is to create a varied, saturated set of statements that address the study’s focus. An example of a focus prompt or for collecting statements is: “With regard to educational ICT projects, my organization has the following challenges OR chances … …” The researcher then edits the statements, assessing for duplicates, splitting statements that contain more than one thought, and removing statements irrelevant to the study, with 80–100 statements optimal (Trochim and McLinden, 2017).

During Step 3, statements are sorted and rated, with participants grouping statements in virtual piles that have similar meanings or relevance, labeling each. Sorting follows several rules—(1) a statement cannot be sorted into multiple piles, (2) more than one pile must be created, and (3) each pile must have more than one statement. Jackson and Trochim (2002) recommend a minimum of 10 sorters, and Rosas and Kane (2012) suggest that 20 to 30 sorters provide the best results. Irrespective of the number of sorters, a fit indicator (stress-value) calculated using the Groupwisdom^TM software (2022) that falls between 0.205 and 0.365 indicates a good fit of the group concept mapping (Rosas and Kane, 2012). After sorting the statements, rating takes place, usually conducted using a Likert-type scale. Participants rate the statements based on the criteria established during Step 1, considering how each statement relates to others.

During Step 4, the Groupwisdom^TM software (2022) is used to analyze the data that participants generated during Step 3. The software uses multidimensional scaling (MDS) and hierarchical cluster analysis to create 2-dimensional point and cluster maps. Points (i.e. statements) proximate on the map are sorted more frequently, and vice versa (Trochim, 1989). When points are grouped, they represent clusters. The researcher uses an iterative process to determine the optimal number of clusters that produce meaningful content. The point and cluster maps are then combined with average ratings of the statements from participants. The result is a layered version of the maps, representing how important participants felt a cluster was. Step 4 thus results in four interpretable visualizations—the point map, the cluster map, the point rating map, and the cluster rating map. During Step 5, these results are interpreted.

Participants were middle and higher managers from the largest public organization in the Netherlands, a police organization, who participated in an education program on leadership and intelligence to better understand leadership in the context of digital transformation. The program took place from March to December 2021, and it was designed to help participants cope with changes from digital transformations. Forty participants (11 females and 29 males) were invited to contribute to the brainstorming portion of the study. Of them, 25 completed the brainstorming task, 22 completed sorting, and 20 completed the rating task. Sixteen completed all tasks—11 middle managers (2 female and 9 male) and 5 higher managers (2 female and 3 male).

The study recruited participants from the leadership course. The study’s structure included online brainstorming, sorting, and rating as separate components. The invitation to participate outlined the data collection procedure and reinforced that participation was voluntary and independent of course completion. Prior to brainstorming, participants were reminded of their voluntary participation, and they were asked for informed consent.

Participants were asked to complete the focus prompt: “Important leadership skills and behaviors in view of the ongoing digitalization of your work are or will be …” They were given two weeks to generate statements based on this prompt, and they answered questions about their management level and gender. After two weeks, the participants had produced 87 statements. The researchers reviewed the statements and removed redundancies, as Trochim (1989) recommends, leaving 61 statements. To ensure a comprehensive set of statements, the researchers added 23 from literature on leadership skills and behaviors during digital transformations ( Appendix), resulting in 94 statements that could be sorted and rated.

Participants were given four weeks to complete the sorting and rating. They began by sorting statements into piles based on their meaning or relevance, and they then rated them on a scale of 1–5, with 5 being very important in comparison to other statements. The researchers excluded data from participants who completed less than 75% of the tasks from the analysis (Schophuizen et al., 2018). Once the participants completed the tasks, the researchers selected the optimal number of cluster solutions that the Groupwisdom^TM software (2022) supported, with the goal of identifying a cluster combination that best represented the data while ensuring internal thematic cluster coherence. The process resulted in five interpretable clusters that were labeled based on their thematic content.

After merging participants’ statements with supplementary statements from the literature, 94 unique statements were obtained. The point map of this set is shown in Figure 2, with each statement assigned a number (see  Appendix for details), representing the relational structure of the statements derived from participants' input (Trochim and McLinden, 2017).

Distances between points were calculated using the GCM tool, considering the bridging values of each point. Bridging values indicate how close statements are to one another, and they can range from zero to one. Statements with lower bridging values were grouped more frequently by participants, and they, therefore, appear proximate on the map, indicating a similar meaning or theme (Trochim and McLinden, 2017). The stress value, calculated by the tool, determines how well the point map visualization matches the data, with values between 0.205 to 0.365 being acceptable (Rosas and Kane, 2012). After 11 iterations, the average stress value was 0.244, suggesting that the map represents the sorted data well.

Hierarchical cluster analysis was based on the point map shown in Figure 2. After considering various alternatives, a five-cluster solution was considered optimal, and after scrutinizing the cluster content, the researchers reassigned four statements to neighboring clusters based on better conceptual fit (see  Appendix). This solution organized the data into interpretable clusters that were labeled based on the leadership skills and behaviors they contained, resulting in five cluster themes—digital technology, working remotely, cooperation, general management, and personal leadership (Figure 3). Statements were assigned a bridging value, combined and averaged for each cluster. When participants frequently grouped statements, the average bridging value of that cluster was lower. Table 1 reports the number of statements in each cluster and the mean bridging value per cluster.

Cluster 4, consisting of leadership skills and behaviors related to personal leadership, had the highest thematic coherence, while Cluster 3, which included skills and behaviors related to cooperation, had the lowest. Cluster 1 comprised ten statements that pertained to leadership skills and behaviors important for managing a digital workforce while working remotely. The cluster included one statement derived from the literature, and bridging values ranged from 0.53 to 0.65 (mean = 0.56), indicating that the cluster was reasonably thematically coherent (1 indicates not coherent at all, 0 indicates perfect coherence). The cluster included both general behaviors and skills, such as ensuring work-life balance and focusing on performance, and behaviors and skills specific to remote working, such as supporting online teams, encouraging remote working, taking opportunities to improve work, and evaluating digital work activities.

Cluster 2 contained 13 statements that described leadership skills and behaviors related to digital technology, referring to knowledge and abilities that relate to using digital technologies effectively when leading and managing. Bridging values for this cluster ranged from 0.44 to 0.76 (mean = 0.58), indicating reasonable thematic coherence and consensus among sorters. Statements in this cluster pertained to working with digital technologies, such as being able to organize information using digital technologies, awareness of both the positive and negative aspects of digital technologies, including monitoring online privacy and security and recognizing the potential and value of digital technologies, such as being aware of the possibilities of informatics and robotics.

Cluster 3 was the smallest, consisting of eight statements that focused on leadership skills and behaviors related to cooperation, or working together with others to achieve a common goal. Bridging values ranged from 0.63 to 0.98 (mean = 0.91), indicating that participants sorted these statements inconsistently. Statements in this cluster included recognizing the benefits of cooperation and investing in external cooperation to enforce digital environments. Cluster 4 was the second largest, comprising 25 statements that discuss leadership skills and behaviors related to the personal development and growth of an individual leader, skills and behaviors that have positive impacts on leadership. The cluster included twelve statements derived from the literature, with bridging values ranging from 0.01 to 0.28 (mean = 0.06), indicating that the cluster had a very coherent theme. Participants largely agreed on the leadership skills and behaviors that are important during digitalization, with examples including authenticity, integrity, tolerance, humbleness, and understanding.

Cluster 5, the largest, included 38 statements on leadership skills and behaviors regarding general management. Statements centered on a leader’s ability to lead and coordinate people, resources, and processes, with ten added from the literature. Bridging values ranged from 0.03 to 0.57 (mean = 0.26), suggesting a coherent cluster. Participants grouped these skills and behaviors because they deal with general management; they support employees, strategic planning, change, decision-making, and communication, including empowering and inspiring people, recognizing opportunities, dealing with change, making quick decisions, and communicating effectively. Four statements were moved from neighboring clusters to this one to improve conceptual fit. Three statements (i.e. fostering interpersonal communication, maintaining personal contact, and creating connections between people) were transferred from Cluster 3, and one (i.e. stimulating) from Cluster 4.

The point rating map was created using the point map (Figure 4) as a basis, with average ratings per statement added. Using a Likert-type scale, participants rated each statement based on its perceived importance in relation to other statements. The colored stacks shown in Figure 4 indicate the ratings, with taller stacks representing greater perceived importance.

The cluster rating map is also derived from participants’ ratings of the statements. It includes average perceived importance ratings of all statements in each cluster (Figure 5). The number of layers in a cluster indicates the degree of importance, and the more layers a cluster has, the more important it is perceived to be. Clusters perceived as most important were 4 and 5, and Clusters 2 and 3 were deemed least important, based on the layers. Importance ratings per cluster were averaged and reported in Table 2, supporting this finding further. The cluster that included skills and behaviors related to digital technologies had the lowest average, and the cluster about general management had the highest.

The pattern match graph (Figure 6) shows the degree of agreement or disagreement among the average ratings per cluster of middle and higher managers. Similarity in the placement of nodes and connections across clusters indicates shared understanding or agreement on aspects of a cluster’s topic. Disparities in the arrangement of nodes or connections between clusters suggest variations in viewpoints or interpretations among the groups. Both groups ranked general management and personal leadership as the most important, and cooperation and digital technology were the least important. Overall, average ratings of the clusters were greater in the higher management group, a difference especially prominent regarding cooperation and digital technology.

To determine if disparities between average ratings were significant, we conducted an independent samples t-test to compare average clusters between middle and higher managers (Table 2). Findings suggest a statistically significant difference in average ratings between the two groups for digital technology (t(24) = 3.37, p = 0.0025) and cooperation (t(14) = 2.20, p = 0.045). In both clusters, mean ratings for the higher managers were greater than those among middle managers.

This study explored critical clusters of leadership skills and behaviors required to manage continued digitalization, investigating whether higher and middle managers perceive the importance of skills differently. The findings suggest five clusters of leadership skills—working remotely, digital technology, cooperation, personal leadership, and general management—with both higher and middle managers ranking general management and personal leadership as the most critical clusters, indicating a shared understanding of the importance of overarching management skills and self-awareness when driving and supporting digital transformation. Both groups rated cooperation and digital technology as least important, perhaps suggesting a potential undervaluation of collaboration and technological proficiency, which are essential in digital contexts. However, a distinction emerged in overall average ratings, with higher managers rating all clusters more favorably than middle managers, a discrepancy particularly pronounced for cooperation and digital technology clusters, suggesting a divergence in how managers at these levels perceive the importance of crucial skills in their roles (Kaiser et al., 2011).

Higher and middle managers' operational and tactical agendas create complexities, especially regarding organizational politics and competition (Obidike, 2019). In such environments, priorities and objectives across departments and stakeholders often conflict, leading to resource allocation, decision-making, and strategic directions becoming points of contention (Kaiser et al., 2011; Van der Velde et al., 1999). These dynamics influence the success of digital initiatives since competing interests result in misalignment and lack of coordination across the organization. Such disparities create friction and impede the implementation of digital strategies. Findings underscore the necessity for strategies that promote alignment and collaboration across managerial levels to address these challenges. Ensuring that higher and middle managers align on the importance of cooperation and technological proficiency is crucial. Higher managers must focus on building organizational frameworks that drive long-term digital adoption (Mergel et al., 2019), while middle managers need tools to manage virtual teams effectively and ensure that operational challenges are addressed (Meijer, 2015). Leadership development programs should cater to these distinct yet complementary needs to support both strategic transformation and operational agility. The divergence in ratings—particularly the higher importance assigned to cooperation and digital technology by higher managers—underscores the differences in focus between strategic and operational responsibilities. Higher managers prioritize long-term, systemic goals, while middle managers emphasize immediate, practical concerns (Spagnoli et al., 2021). These findings align with Mintzberg’s managerial roles framework, which highlights how strategic roles such as resource allocation and entrepreneurship dominate higher management, while middle management is more focused on leadership, monitoring, and disturbance handling to address workforce-level challenges (Mintzberg, 1973). Cluster 5 (general management) reflects this broad spectrum of competencies, analogous to Mintzberg’s (1973) extensive list of managerial roles spanning both strategic and operational priorities. While this breadth demonstrates the foundational importance of general management, its size raises questions about whether the cluster should be further refined to benefit more targeted leadership development.

To bridge these gaps, additional analyses using Go-zone plots can provide valuable insights (Bar and Mentch, 2017). Such plots serve as pattern match figures that reveal underlying structures of participants' perceptions, identifying common themes, patterns, and areas of contention. They facilitate decision-making and planning based on consensus and areas of divergence (McLinden, 2017). The go-zone plot in Figure 7, representing the statistically significant different clusters of digital technology and cooperation, serves as a starting point to determine where common ground lies regarding important leadership skills and behaviors. It depicts statements in a quadrant map where the x-axis represents middle managers and the y-axis higher managers, with the upper right quadrant showing statements that score high in both groups.

Group learning (Wilson et al., 2007) can play a crucial role in addressing perceptual gaps between higher and middle managers. By facilitating collaborative learning environments, organizations can enhance mutual understanding and align strategies across different managerial levels (Virolainen et al., 2021). This collective approach contributes to targeted leadership development programs while fostering improved communication channels and integrating project teams. Ultimately, group learning could cultivate a unified perspective that accelerates the adoption of digital initiatives, ensuring that all levels of management are engaged in the transformation process (see also, Virolainen et al., 2021).

This study’s participants, comprising higher and middle managers from the largest public organization in the Netherlands, provide context-specific insights for designing tailored leadership development programs (Lawrason et al., 2023). Such programs should address the identified perception gaps, ensuring both management levels recognize the importance of critical leadership clusters. By fostering a culture of mutual understanding and collaboration, organizations can enhance their ability to implement digital initiatives (Henderikx and Stoffers, 2023). The differing perceptions of essential leadership skills and behaviors between higher and middle management underscore the necessity for these tailored programs, which should include customized training and mentorship opportunities designed to meet the unique needs of each management level. The government’s role in the public sector is critical; it can establish policy frameworks that promote digital literacy, allocate resources for skill development, and forge partnerships both within the public sector and with the private sector to integrate best practices in digital leadership. Engaging various stakeholders through workshops and cross-departmental teams can facilitate productive discussions that align diverse perspectives on the skills required for effective digital management. Additionally, organizations should conduct comprehensive skills gap analyses, implement continuous feedback loops to assess the effectiveness of leadership programs, and involve external experts specializing in digital transformation. This comprehensive approach can bridge the skill perception gap and foster more effective navigation of the complexities associated with digitalization. Recognizing that the differing perspectives of higher and middle management are valuable—representing two distinct realities that collectively enable successful transformation—is essential.

While bridging the gap for mutual understanding is critical, it’s also important to appreciate the unique perspectives and experiences of both management layers. Future higher management will emerge from middle management; however, selecting them based on alignment with departing executives' views should be avoided. They should bring a shift in priorities, shaped by evolving government and societal challenges. Recognizing this dynamic is crucial for preparing middle managers for executive roles that balance continuity with adaptability, particularly as public-sector priorities change over time (Bresciani et al., 2021).

Embracing this complexity ultimately strengthens the organization’s capacity to adapt and thrive in an increasingly digital landscape.

Although it offers in-depth insights, this study’s qualitative design was constrained by some limitations, such as sample size (Alsaawi, 2014), subjective data interpretation (Sauerborn, 2019), and reliance on single-source data collection (Lewis, 2009). Also, this study did not include gender-based analyses; the presence of fewer female participants reflects a broader trend in public-sector leadership. Exploring whether men and women prioritize different leadership skills during digital transformations would enrich understanding and inform gender-sensitive leadership programs. For example, prior research suggests that female leaders often emphasize collaborative approaches and empathy, traits potentially critical for fostering cooperation and driving digital initiatives (Hermkens, 2021). Future research should aim to bridge this knowledge gap by examining gendered leadership perceptions in digital contexts. Despite these limitations, the study contributes to extant leadership literature by highlighting disparate perceptions between higher and middle managers (Van der Velde et al., 1999).

This study highlights critical clusters of leadership skills and behaviors essential for navigating two aspects of digitalization: leading in a digitally transformed environment and leading the transformation process itself. The findings reveal disparities in the perceptions of higher and middle managers regarding these skills, emphasizing the need for tailored leadership development initiatives that bridge strategic and operational priorities, fostering alignment across managerial levels (Mergel et al., 2019; Meijer, 2015). These insights underscore the importance of tailoring leadership development to address varying priorities and perspectives across managerial levels, ensuring cohesive approaches to digital transformations. The originality of this study lies in its participant-driven approach to investigating perceptions of higher and middle managers regarding leadership skills and behaviors during digital transformations. First, to address such disparities, this study suggests new topics for both research and practice, providing insights for organizations that wish to thrive in the digital age. These insights inform not only leadership development, but strategic alignment of digital transformation, contributing to cohesive and effective organizational change.³