The Limitations of Self-Managed Teams

You no doubt have heard about Zappo’s, the on-line shoe retailer that offers $2000 to new employees who decide after their initial training that they no longer wish to continue with the company. A couple of years ago, Zappo’s founder, Tony Hsieh, decided to introduce a management concept called holacracy (a variation of the self-managed team concept) and offered all employees the choice of staying with the company and embracing this new concept or taking a generous severance package. Eventually, by the time the deadline had passed, 260 of the company’s roughly 1500 employees (about 18%) decided to take the offer – which was on average about 5.5 months’ worth of salary.

Zappo’s specific implementation of holacracy was developed by Brian Robertson, a former programmer, and is perhaps a more structured and extreme variation of the concept of self-managed teams.  Zappo’s is not the first company to experiment with some form of self-managed teams; Thorell (2013) claims to have identified 18 “bossless” companies in the U.S. (not including Zappo’s), with 6 of them operating internationally.

Another company, W. L. Gore, has become famous not only for its revolutionary fabric but also for its commitment to self-managed teams. Here is an excerpt from its web site:

“How we work at Gore sets up apart. Since Bill Gore founded the company in 1958, Gore has been a team-based, flat lattice organization that fosters personal initiative. There are no traditional organizational charts, no chains of command, nor predetermined channels of communication.

“Instead, we communicate directly with each other and are accountable to fellow members of our multi-disciplined team. We encourage hands-on innovation, involving those closest to a project in decision making. Teams organize around opportunities and leaders emerge. This unique kind of corporate structure has proven to be a significant contributor to associate satisfaction and retention.”

The distinguishing feature of such self-managed teams is the absence of any formal hierarchy. This means there are no formal titles or promotions, and no one has a “boss.” Team members negotiate their responsibilities with each other, employees rank each other to help with pay decisions, and team members can hire and fire.  These applications of self-managed teams are consistent with Carson et al.’s (2007) definition: those in which team members have greater responsibility for setting their own goals, monitoring their own progress, and making their own decisions than do team members in manager-led teams.

Note that self-managed teams are not, strictly speaking, leaderless. Rather, leadership is “shared” or distributed. In their meta-analysis on the impact of “shared” leadership on team effectiveness, Wang et al. (2014) conclude that the effects of this type of leadership is stronger for attitudinal outcomes and group processes than on team performance.

Actually, the majority of work organizations not only in the U.S. but globally are based on hierarchical principles. Some have tried to experiment with flat organizations and have not had much success. Even Google learned this the hard way, when its founders Larry Page and Sergey Brin flattened the organization early on but pulled the plug a few months later when they found that this was just not working.

According to Galinsky and Schweitzer (2015), hierarchy is the most dominant form of social organization across all species: “Hierarchy helps people know who does what, when, and how. These rules promote efficient interactions by setting clear expectations for the behaviors of people of different ranks. Essentially, hierarchy facilitates social interactions by simplifying them.” (p. 67)

Kenney and Anderson (2012) have summarized the literature on status hierarchies in groups, and they point to two theories of hierarchy: the functionalist and the dominant theories, and propose a third one, what they call the micropolitics theory. In brief, the functionalist theory says that hierarchies form to help groups make decisions more efficiently. They give some individuals who are seen to be the most competent more control over decisions. According to the dominance theory, hierarchies form as a result of competition and the assertiveness of some members over others. While the research shows overwhelming evidence for the functionalist theory over the dominance theory, the authors propose a blending of the two: that hierarchies form as a result of who can help the group most and also through individuals’ desire for status. Regardless of the evidence for each of these theories, the reality is that most social and work groups establish hierarchies early on, and they are relatively difficult to eliminate.

In my view, four factors need to be taken into account for any organization interested in moving towards self-managed teams: cultural preferences, individual differences, organizational context, and the nature of the work or task. First, cultural preferences. As Hofstede and others have demonstrated over the years, cultures that are high in uncertainty avoidance and high in power distance in particular have a strong preference for hierarchy. Several years ago, I was coaching a Taiwanese executive who was the general manager of a global company’s subsidiary. His company had just introduced the matrix form of organization, and Ron (not his real name) commented to me: “This is difficult. Most Asians want to know ‘who is my boss?’”

In a recent study, Herbert et al. (2014) found that perceptions of shared leadership varied depending on the collectivistic orientation of the participants in their study (over 350 members of various virtual project teams). And in a survey of over 200 Mexican executives, Nicholls et al. (1999) also found numerous challenges in the implementation of self-managed teams in the Mexican work setting. As demonstrated by Hofstede and others, Mexican cultural values are closely related to collectivism, high power distance, and high uncertainty avoidance. While a collectivistic orientation might help provide a positive reaction to self-managed teams, the last two certainly do not, as their interview findings suggest. Anicich et al. (2014) have been studying the effects of hierarchy both structurally and culturally. In their studies of expert mountain climbers from 56 countries on over 5,000 expeditions, they found that expeditions from more hierarchical countries had more climbers reaching the summit, but also more climbers dying along the way. These findings illustrate both the benefits and drawbacks of hierarchy, which can create efficiencies but, by suppressing participation and voice, can be dysfunctional.

Second, individual differences. People will vary in their need for structure and their own motivational level, as well as in their testosterone levels. I would predict that individuals who tend to have a strong desire to be open to experience (one of the personality characteristics of the Big Five theory of personality) are more likely to be comfortable with such self-managed teams. Furthermore, many studies have shown that high-testosterone individuals are dominant and like being in high-power positions. By implication, such individuals like to be in charge and prefer hierarchy. In fact, as Galinksy and Schweitzer (2015) and others have written, two baboons that have high testosterone levels become very competitive and when testosterone levels vary, the baboon with the lower testosterone will walk away. Similar findings have been found in research on other primates and animals. Perhaps this is why co-CEOs are rare and when they do happen, they do not last long. It is rare for two alpha dogs to cooperate over a long period of time! As an aside, Galinksy and Schweitzer (2015) suggest a simple test to determine a rough estimate of your testosterone level: compare your ring finger with your index finger. The longer the ring finger is relative to the index finger, the more you were exposed to greater levels of testosterone in the mother’s womb – and this exposure is an indicator of your testosterone.

Third, the organizational context. Moving from an authoritarian corporate culture to one that is self-managing is not an overnight transition, as anyone knows who has worked in companies trying to create large-scale change. In fact, there is something ironic in a leader who dictates that henceforth his or her company/division/group will implement self-managing teams. In addition, because there will inevitably be a period of confusion and uncertainty when this concept is implemented, it is generally not a good idea to introduce it when an organization is going through a number of changes at the same time (e.g., introducing a new business model while restructuring and downsizing).  

One of the successful examples of a self-managed team is the Orpheus Chamber Orchestra, an ensemble of musicians that has operated without a conductor for many years. The orchestra has been written about and studied over the years (e.g., Vredenburgh and He, 2003) and it has certainly thrived. Yet if you examine the context, several features are notable. First, it has only about 25 musicians, since it is a chamber orchestra, not a full orchestra. Second, it has established a culture of collaboration which is congruent with the self-managed team approach. Third, members have self-selected into the orchestra so that its musicians are motivated by self-leadership and reciprocal influence. The orchestra is, however, NOT leaderless; rather, leadership roles rotate. Here is one of their principles and a brief explanation: Working in shared leadership. Every musician has the opportunity to play leadership roles such as leading rehearsals or directing the performance of a new musical composition. The decision rests on the group; the leader is chosen according to expertise, strengths and interests (Luc, 2011).”

Fourth, the nature of the work or task. Self-managed teams seem to work best when members are working on tasks that are complex and require high levels of collaboration and information sharing. Ronay et al. (2012) performed a series of experiments in which they assigned participants to different groups working on tasks that were either high in procedural interdependence or low in procedural interdependence. Participants’ power levels were “primed” so that there were three different types of groups: groups of 3 high-power participants, groups of 3 low-power participants, and groups of 1 high-power, 1 low-power and 1 baseline participant. The mixed-power groups were significantly more productive than the other two groups but only for the tasks high in procedural interdependence. In a second experiment, they created different types of groups (groups of high-testosterone participants, groups of low-testosterone participants, and a mix of high-, low- and average-testosterone participants) and had them working on the interdependent tasks. The mixed groups were significantly more productive than the other two groups, with the high-testosterone groups experiencing much more intragroup conflict than the other two types of groups. The researchers conclude that “… the functional benefits of hierarchy are most pronounced under conditions of high procedural interdependence …” and “…intragroup conflict mediated the performance decrements for the high-testosterone groups, but not the low-testosterone groups.” (p. 675)

The intent of organizations promoting self-managing teams is certainly admirable. Too many organizations have created cultures where people are afraid to speak up, and where ideas (especially from the rank-and-file) are not listened to. The effects of too much hierarchy, combined with the presence of high-testosterone individuals, many of whom are also narcissistic, can be damaging for teams and organizations that not only need to be high performing, but also need to be innovative and resilient. Giving employees more of a voice and empowering them is in principle a great idea. And moving away from the command-and-control model to a more collaborative model will resonate with many employees today.

For organizations interested in making the move towards self-managing teams and less hierarchy, here are three pieces of advice. First, assess your organization’s readiness for this change (more on this in a subsequent blog). Second, do some experimenting and piloting, while making sure you involve organizational members and get their feedback. Third, allow for transitions, keeping in mind that the choice is not necessarily a binary one between an authoritarian, hierarchical model and a self-managed model – but may well lie in variations on this continuum.


Anderson, A., and J. Kennedy. (2012). Micropolitics: A New Model of Status Hierarchies in Teams, in Margaret A. Neale, Elizabeth A. Mannix (eds.) Looking Back, Moving Forward: A Review of Group and Team-Based Research (Research on Managing Groups and Teams, Volume 15.) Emerald Group Publishing Limited, 49 – 80.

Anicich, E.M., Swaab, R.I., & Galinsky, A.D. (2014) Hierarchical cultural values predict success and mortality in high-stakes teams. Proceedings of the National Academy of Sciences. doi: 10.1073/pnas.1408800112

Carson, J. et al. (2007). Shared Leadership in Teams: An Investigation of Antecedent Conditions and Performance. Academy of Management Journal, 50 (50: 1217-1234.

Galinsky, A. and Schweitzer, M. (2015). Friend & Foe. New York: Crown Business.

Garvin, D. (2013). How Google Sold Its Engineers on Management. Harvard Business Review.

Luc, E. (2011). The 8 Leadership Principles of Orpheus, the Conductor-less Chamber Orchestra.

Nicholls, C. et al. (1999). Taking Self-Managed Teams to Mexico. Academy of Management Executive, 12 (2): 15-25.

Ronay, R. et al. (2012). The Path to Glory Is Paved with Hierarchy: When Hierarchical Differentiation Increases Group Effectiveness. Psychological Science, 23 (6): 669-677.

Thorell, L. (2013). How Many Bossless Companies Exist Today?

Vredenburgh, D., and I. He. (2003). Leadership Lessons from a Conductorless Orchestra.Business Horizons, September-October, 19-24.

Wang, D. et al. (2014). A Meta-Analysis of Shared Leadership and Team Effectiveness.Journal of Applied Psychology, 99 (2): 181-198.