Google learned the hard way why hierarchical structure is important

Saerom (Ronnie) Lee, Assistant Professor of Management at the Wharton School, University of Pennsylvania, argues countless examples of flat structures failing highlight the need for hierarchical structures in business.

Why have flat organisational structures become ‘trendy’ in recent years?
 
Actually, flat organisational structures are not totally new. These structures have been “trendy” since at least the early 20th century. For example, Lyndall Urwick wrote in a Harvard Business Review article (entitled “The manager’s span of control”) in the 1950s that: “At present, there are many in favour of the flat setup.” Even then, many academics and practitioners argued that “authoritarian” hierarchical structures are outmoded and will be supplanted by flat “egalitarian” structures.
 
Despite the pervasiveness of hierarchical structures to this day, this argument for (or fascination with) flat hierarchies has persisted over time. Recently, these flat structures have received attention because of companies like Valve, Github, Medium, Morningstar, Zappos, Oticon, and Supercell. For their egalitarian ideal, these companies have been praised by popular media outlets and examined by numerous scholars. And, prior studies have suggested that modern workers (in particular, women) would strongly prefer such an egalitarian structure over tight managerial supervision and a strict chain of command.
 
However, in a recent study, my colleagues and I find quite the opposite. That is, using two large-scale labour market experiments, we show that a flat hierarchy does not have a significant effect on the number of applicants but, in fact, decreases the proportion of female applicants. These findings imply that flat hierarchies could inadvertently curtail employee diversity and, ultimately, be detrimental to firm performance and innovation.
 
What advantages are flat structures seen to hold over more hierarchical structures?
 
Prior research suggests that flat hierarchies can have various advantages over more hierarchical structures. Most notably, these flat organisational structures could create a more egalitarian work environment that provides employees with more autonomy and flexible work arrangements. This work environment can motivate and encourage employees to work on tasks that they find interesting and meaningful, to explore and generate more creative ideas, to share these ideas and voice their opinions, and to enjoy a work-life balance. By motivating employees as such, companies may foster creativity and innovation and more quickly adapt to their rapidly changing environments. Although my research partially supports these arguments, it suggests that flat hierarchies could hinder coordination and be detrimental to firm performance and survival.
 
How do flat structures avoid the emergence of unofficial hierarchies over time?
 
In theory, flat structures could avoid the emergence of unofficial status hierarchies by defining a clear set of rules, responsibilities, procedures, norms, and organisational culture. However, this is easier said than done. When this hierarchical structure is not formally defined and agreed upon by its constituents (as in flat hierarchies), an informal status structure will emerge over time through social interactions (such as power contestations, politicking, and conflicts).
 
For example, a former employee of Valve (the video game company known for its flat hierarchy) noted that: “The one thing I found out the hard way is that there is actually a hidden layer of powerful management structure in the company and it felt a lot like high school. There are popular kids that have acquired power in the company, then there are the trouble makers who actually want to make a difference”.

Thus, this unofficial status hierarchy seems somewhat inevitable in reality, but the degree to which this informal structure is dysfunctional would vary by individual (power-seeking behaviour), organisational (rules and procedures), and institutional (national culture based on age and seniority) attributes.
 
How feasible are flatter structures for larger organisations such as Google? Is there a template to follow or is it dependent on the goals of individual organisations?
 
Whether flatter hierarchies are feasible for larger organisations will depend on various organisational characteristics, such as their industry, their employees, and the operating procedures. Using a unique combination of these characteristics, the companies I talk about earlier have found ways for flatter organisational structures to work for their large organisations. However, these companies are a rarity and, given the complex interactions and causal ambiguity in these characteristics, it remains unclear what the ‘secret formula’ is.
 
Although many large companies have experimented with flat organisational structures, most of these experiments have unfortunately failed, as discussed in my study. In fact, a well-known example is Google. In 2001, Larry Page, the co-founder and then CEO of Google, decided to remove middle managers, assuming that they are creating layers of unnecessary bureaucracy. Eliminating these middle managers, ultimately, led more than a hundred engineers to report directly to a single executive. As a result of this exponential increase in their span of control, these executives were overwhelmed with coordinating these engineers, and these engineers received little managerial guidance and lost direction. Thus, Google learned the hard way why hierarchical structure is important for coordination and ended up reinstalling the middle management.
 
How crucial is technology in ensuring that flat organisational structures can be established and consolidated?
 
Recent studies of flattening firms by Julie Wulf (Harvard Business School) and colleagues have found that information technologies can facilitate information flow between employees and may thus partly fill in the coordination role of middle managers. By reducing the layers of middle managers, these technologies may thus help companies cut costs (especially the wages of middle managers) and become more flexible and responsive to their environment. However, substituting middle managers with IT can increase the span of control (the number each manager supervises) and thus put more burden, or “cognitive overload” on the top management team.
 
In addition, if IT increases the complexity of the company’s underlying business or products or if they disproportionately increase the amount of information that the company needs to process, the top management team could face even more challenges in coordination and decision-making.

In reality, the answer depends on how IT affects the company’s information processing demand (how much information the company needs to process) and information processing capability (how efficiently the company can process information).