Interest in “flattening” firms—reducing the number of layers between the CEO and the lowest level on the organization chart—began more than 40 years ago (e.g., Toffler 1981; Peters and Waterman 1982).

It probably began in part as a response to the conglomerates that had been built up during the 1960s and their lagging performance in the 1970s. Though an increasingly distant memory, these behemoths combined hundreds of businesses, remotely related if at all. ITT, for example, initially a telephone and telegraph company, acquired divisions that did everything from operating hotels to offering insurance to baking bread.

Flattening these firms usually involved not just reducing the number of managerial layers but also restructuring them. Firms would emerge from these processes not just flatter but also smaller and more focused.

Most of these conglomerates have long since disappeared, but managerial interest in flatter firms has persisted. Markus Reitzig’s essay, and even more so the book upon which that essay has been based (Reitzig 2022), provides a how-to manual for these managers. Importantly, however, Reitzig does not stop at how-to, he also answers the questions of when and why managers might want to streamline their hierarchies.

The proponents of flatter firms often present them almost as getting something for nothing. Flat firms continue to produce the same goods and services. But, with fewer layers of middle managers, they presumably produce those things at less cost, more efficiently. Why, then, have all firms not adopted these flat structures?

Reitzig (2022) points out that flat organizational structures do not necessarily benefit all firms. Middle managers matter. They mentor, monitor, and motivate their employees. They determine who needs to do what. They resolve conflicts between employees. They filter and communicate relevant information both from their own bosses to those they supervise and vice versa. All of these activities contribute to organizations operating efficiently and effectively.

Flatter firms, those with fewer of these middle managers, Reitzig argues, appear most useful to meeting three goals: (1) Flat structures stimulate innovation, by allowing employees to interact with more diverse alters (for recent empirical evidence, see Lee 2022). (2) They enable faster adaptation. Hierarchy helps to ensure accountability and reliability but simultaneously retards the speed at which the organization can respond to changes in technology and customer preferences (Hannan and Freeman 1984). (3) Flat structures, by providing people with more interesting and satisfying jobs, may also help to recruit and retain employees.

Complementary choices

For me, one of the great strengths in Reitzig’s analysis comes from his attention to complementary choices. Managers, he maintains, cannot simply eliminate managerial layers without simultaneously changing other dimensions of the organization design.

For example, flattening the firm usually also requires a change in compensation schemes. Middle managers help to monitor and motivate employees. As they become fewer in number, the span of control—the number of employees they oversee—increases. Managers have less time to monitor each report. To mitigate the agency problems that might ensue, firms that succeed at becoming flatter usually adopt compensation practices that more closely align the financial interests of employees with those of the firm (Reitzig 2022).

Removing layers of hierarchy also typically means devolving decisions. Middle managers have less time to direct those they supervise. To prevent them from becoming a bottleneck in organizational processes, successful flattening therefore generally means giving employees the ability and authority to act more autonomously (Reitzig 2022).

The complementary choices do not end there. Flattening the firm may require a reorganization of reporting lines, an expansion of routines, and even adjustments in the corporate culture. Indeed, Reitzig (2022) discusses interactions between the depth of the corporate hierarchy and nearly every other aspect of organization design.

In his how-to format, Reitzig presents these complementary choices as a practical matter. But I believe that his arguments raise a larger point that has been under appreciated in the literature on organization design.

For more than 20 years, I have been using the People, Architecture, Routines, and Culture (PARC) Framework to teach organization design (for an excellent review, see Saloner et al. 2001). Most discussions of the framework treat it as a set of semi-independent policies. Managers can pick and choose different policies to tune the organization to some optimal balance of coordination and individual incentives.

Reitzig’s discussion of these choices in the context of flattening the firm, however, highlights the fact that they are not independent. They are complementary. Rather than having a near-infinite set of feasible combinations available to them, managers instead choose between a few clusters of coherent organizational design choices.

This view of organization design as a set of complementary choices interestingly brings it in line with current thought in other closely-related areas. Human resource management, for example, has increasingly been focused on clusters of policies that maximize performance (e.g., MacDuffie 1995; Baron et al. 2001). Strategy has similarly come to view competitive advantage as being a function of adopting sets of coherent choices (e.g., Milgrom and Roberts 1990; Rivkin 2000; Porter and Siggelkow 2008).

The fact that flat organizational designs fit best with firms trying to emphasize innovation and adaptability, moreover, further suggests that these complementary design choices may also complement specific strategic positions. Form follows function.

Employee effort

Although I appreciate the book and the essay and I recommend them without reservation, I find myself disagreeing with one of the claims. Reitzig (2022) argues that flat firms require more of their employees, not just in the sense of being proactive but also in the effort that they exert. They therefore must recruit people willing to work harder.

At first blush, this conclusion seems straightforward. Fewer middle managers means that others must now cover the responsibilities of those eliminated, in addition to everything that they had already been doing. Each employee must do more.

However, I believe that this conclusion rests on two implicit assumptions: first, that flattening leads to fewer employees; second, a conservation of activities and of the time required for them. In other words, the flattened firm must still do everything that had been done before but with fewer people.

But those assumptions may not hold true. As firms flatten their hierarchies, they may reduce their ranks of middle managers but expand in other areas, such as in sales or in client relationship management. The total number of employees therefore may remain unchanged.

Even if the headcount does decline, if firms restructure themselves in the ways that Reitzig advises, many of the activities the middle managers did will disappear. Incentives will substitute for monitoring. Routines and modularization will reduce the need for resolving conflicts. Shortening the number of nodes between the sources and destinations of information will eliminate a friction in its movement. Despite having fewer people, a flatter firm may not need them to exert any extra effort.

Although the flatter firm might operate more efficiently, these changes do not come without cost. Monitoring ensures not just that employees exert effort but also quality control. Conflict resolution capacity allows the firm to engage reliably in more complex and more tightly-integrated production processes. Lean and agile may mean being more fragile.

The pyramid principle

Finally, although it falls somewhat outside the scope of Reitzig’s research, I feel as though his essay begs a question: Why do managers have a perennial interest in flattening their firms? If managers understand the benefits of being flat, why have more of them not built their firms flat from the beginning?

We need a theory by which firms systematically end up with more hierarchy than ideal.Footnote 1 Let me offer one, which I will call the pyramid principle.

Assume that each bottom-level employee in an organization can produce a certain amount of whatever good or service (e.g., Williamson 1967; Keren and Levhari 1979).

Let us further assume that firms begin with a fixed span of control (at every level). As one might surmise from Reitzig’s discussion, this span of control stems from a large number of interdependent choices. Changing the span of control therefore comes at considerable cost.

In terms of the imagery of a pyramid, the area covered by its base represents the total production of the firm. Span of control essentially defines the steepness of the sides. As the firm produces more – as its base becomes larger – its height must rise if the angle of its sides remains constant.

However, to understand why firms might end up with too many layers, we need two additional assumptions. First, assume that adopting a larger span of control costs more. Reitzig (2022) provides justification for this idea (see also Lee 2022): Developing routines and investing in technology that might allow managers to supervise more people incurs a semi-fixed cost (or at least the marginal costs decrease rapidly with scale).

In essence, then, the benefits of a broad span of control increase at an increasing rate with firm size. But its costs remain relatively fixed. As a result, smaller organizations prefer smaller spans of control.

If we also assume that firms begin small, then we have a world in which firms initially choose small spans of control. They persist with these small spans because the interdependent design choices mean that any change would come at considerable cost. As they grow, their hierarchies therefore become ever more inefficient.

At some point, the benefits of flattening the firm will outweigh the costs of the change. Reitzig (2022) then instructs managers as to how to increase the span of control. In the meantime, however, the firm has more than the (static) equilibrium number of layers.

This simple theory suggests some directions for future research on organizational hierarchies. For example, far from being flat, startups should have a hierarchical height proportional to their logged headcount. They should grow taller as they add employees.

The slope of this relationship between height and headcount, however, should flatten out eventually, as firms begin to invest in the routines and technology that enable larger spans of control. Interestingly, Lee (2022) finds just such a pattern among startups in the video game industry, with the shift in the slope falling someplace around 200 employees.

Startups might also vary in their early spans of control. Being richer in resources, for example, may permit some to invest in complementary technology earlier. Or, recruiting early employees from large, established firms might allow startups to import effective routines for coordination. Those able to adopt larger early spans of control may then enjoy a competitive advantage as they begin to scale (e.g., Lee 2022).

Get Better at Flatter provides both a nuanced perspective on organization design and much-needed guidance for managers. But it also reminds me of how little empirical attention has been given to the topic. I hope that the book will inspire not just managers but also researchers to give greater attention to the pros and cons of organizational hierarchy.