Span of Control and Span of Attention

In our paper, Span of Control and Span of Attention, which was recently made publicly available on SSRN, we use novel data to better understand the role of the CEO and the relationship to the executive team as represented by the CEO’s span of control. We collect detailed time use information for a large sample of CEOs and use it to characterize how CEOs allocate their time. We compare how this new and more comprehensive measure—span of attention—is related to the more traditional notion of span of control.

Overall, our findings suggest that the structure of the executive team—both the size and the composition—is correlated with CEO attention. We find a positive complementarity between the size of the executive team (span of control) and the attention the CEO allocates to internal interactions. Our findings suggest that CEOs invest in a “team” model of interaction in larger teams: they spend more time internally, in planned meetings with multiple participants across functional areas. Taken together, the patterns of “what CEOs do” are consistent with top management team (TMT) predictions that achieving behavioral integration is harder in larger teams, especially in large organizations. Our results suggest, that in these settings, CEOs allocate attention to increase the cohesion among team members. Our findings are consistent with “what CEOs say” about the rationale behind increasing their span of control: to get more involved in internal operations (Neilson and Wulf, 2012). Finally, our analysis confirms that certain positions act as delegates: CEOs allocate less time to internal interactions in the presence of a COO.

Our analysis demonstrates the importance of collecting data on both the formal and informal aspects of organizations. Time allocation of executives, in conjunction with the size and composition of the team, allows for a much richer understanding of the role of the CEO and senior positions, and in particular, the behavior of the executive team. In a survey of TMT research, Finkelstein, Hambrick and Cannella (2009) list three issues that “if solved, could move the area significantly forward.” We believe our measures of time use or span of attention have the potential to address two of the three issues. The first is how to define the boundaries of the TMT in an objective way. While managers may be formal members of the executive team based on title or position in the hierarchy, interactions may reveal the true boundaries of the team. The second issue is how to study the role of the CEO within the TMT. Allocation of attention or time use goes far beyond any existing measure that captures CEO impact on and involvement with the TMT. For instance, time allocation can identify the crucial distinction as to whether a CEO is externally or internally focused.

Although our study presents several limitations, which we discuss in the paper, we believe that a more systematic study of managerial attention over a large sample of executives could bring significant benefits to scholars in both management and economics. First, looking at interactions in conjunction with organizational structure may provide the basis for more realistic formal models of organizations. More generally, time use analysis allows scholars to shift from the focus on individual managers to analyzing teams and complementarities inherent in human capital. At the same time, this study also suggests the importance of refining definitions and sharpening the empirical predictions of top management team models prevalent in the management literature. For example, our findings suggest that it would be beneficial to move beyond the use of simple demographic information of top team members, and consider in more detail the structural aspects of the executive team.

Practitioners may also benefit from time use data for several reasons. First, basic calendar analysis can be used as a tool for assessment for individual managers to evaluate their own patterns of time allocation. In the debrief with the participants of the executive education program at Harvard Business School, for example, we found that the executives’ assessment of how they spend their time was quite different from actual records. Executives also find it fascinating to compare their time allocation to that of other peers in the same industry, as a way to gauge differences in their respective managerial styles. Finally, time use analysis enables senior managers to evaluate whether they are allocating time in a manner that is consistent with firm strategy and priorities critical to the implementation of strategy.

Ultimately, we argue that it is critical to complement traditional notions of organizational structure and team characteristics—span of control—with rich data on CEO interactions—span of attention. Although the collection of time use data for large sample of individuals presents clear methodological challenges, our experience across several countries and the availability of electronic calendars convince us of the feasibility of this research agenda. Most importantly, we believe that this strand of research could facilitate the establishment of a much-needed connection between the formal models of organizational structure in economics and the richness of the theories and empirical findings of the management literature.

Obviously, this paper is just an initial step in this direction, and there are many ways in which the time use data can be improved, suggesting several directions for future research. First, by collecting information on the purpose of the meetings, we could distinguish between interactions that primarily facilitate information exchange versus interactions that involve decision-making and the relationship between objectives and top team structure. Second, by collecting data on the firm’s strategy and scope of businesses, we could explore how interactions and team structure relate to the interdependence of tasks. For example, the role of the CEO might be very different in firms that operate in related businesses versus diversified firms, and we could characterize the CEO’s role by the nature of interactions and relate it to team structure. Also, it would be interesting to analyze smaller groups of teams using our span of attention measure since all team members are not involved in every decision. Lastly, span of control complemented by span of attention will allow us to explore the relationship between organizational structure, management interactions, firm strategy and performance. We leave these topics for further research.

The full paper is available for download here.