The romance of the bossless company collides with a harder truth: when pressure rises, someone still has to decide. Leaders torn between agility and control face a false binary. Flattening removes visible bosses but often installs hidden ones. The better move is to give your network a spine, a minimal and explicit structure that preserves speed without smothering initiative.
What flatness promises, and what it cannot deliver alone
Self-managing organizations decentralize authority formally and systemically, and their defining feature is the elimination of the traditional manager–subordinate relationship (Lee and Edmondson, 2017). The appeal is clear. Classic managerial hierarchies excel at reliable execution in stable environments, yet they tend to inhibit complex, cross-boundary problem solving and reinforce status differences that stifle development (Lee and Edmondson, 2017). No wonder the debate became polarized, with enthusiasts celebrating bossless utopias and skeptics dismissing the whole project as naïve (Bernstein et al., 2016).
Yet polarization is a poor guide for design. Evidence from firms that tried radical flatness shows how networks wobble without a clear backbone. Flat or near-bossless organizations often work under stable conditions. In dynamic and time-sensitive contexts, many struggle with innovation, scale, and longevity. Formal hierarchy is frequently replaced by informal hierarchy rather than eliminated (Foss and Klein, 2022). In other words, the authority you banish from the org chart reappears in back channels.
The cautionary case of Valve
Valve’s widely cited flat structure illustrates the risks. By 2013 employees described a period of creative drift as “The Wilderness,” a stall that contradicted the promise of frictionless self-organization. The development of Half-Life: Alyx ultimately required management intervention despite the firm’s bossless model (Foss and Klein, 2022). Valve also avoided building dedicated customer support and marketing departments. That choice, consistent with an anti-hierarchy posture, created predictable blind spots in functions that coordinate at scale (Foss and Klein, 2022). The lesson is not that networks fail. It is that complex, time-critical work imposes coordination demands that ad hoc influence cannot reliably meet.
Zappos’s arc: from holacracy to a market inside the company
Zappos became the most visible large-scale adopter of holacracy by 2016, the year it drew intense attention as a live experiment in self-organization (Bernstein et al., 2016). The path was bumpy. In March 2015, the CEO offered a buyout tied to full commitment to holacracy, and 18 percent of employees took it (Groth, 2020). Over time, Zappos stripped out the system’s bureaucratic meetings and brought managers back. By March 2017 it replaced holacracy with an internal market in which teams operated like small businesses with their own P&Ls and transacted with one another at prices, a shift prompted by “big challenges” in business metrics and a critique that holacracy was too inward-focused (Groth, 2020). The pivot did not abandon autonomy. It reframed it around explicit commercial commitments and clearer accountability.
Taken together, Valve and Zappos underscore a practical point. If you try to swap hierarchy for a pure network, hidden hierarchies congeal and coordination gaps open. If you give the network a spine, whether through a defined escalation path or an internal market with real decision rights, you can keep initiative alive while avoiding paralysis.
What the spine looks like in practice
The research record does not settle the question of a single best structure. It does, however, point to a contingent answer. Near-bossless organizations are neither generally superior nor common. Managers matter as functional actors who design, organize, encourage, and enforce. There are two distinguishable kinds of managerial authority: one rooted in the right to hire, fire, and instruct, and another rooted in shaping structure and alignment through leadership and consensus (strategy+business, 2022). That distinction is a design tool. Use it deliberately instead of letting it emerge by accident.
Start with explicit decision rights for time-sensitive work. Identify the handful of commitments that cannot wait for consensus. Name who can decide, on what inputs, by when. Keep the list short to avoid recreating the bureaucracy you left behind. This is not a reversion to command and control. It is an admission that some bets need a clear final call, especially when interdependencies are tight. The evidence that flat firms often end up with informal bosses suggests you are better off making those roles explicit with guardrails (Foss and Klein, 2022).
Pair that spine with spaces where the network does its best work. Self-managing structures shine when teams face novel, cross-functional problems that benefit from distributed judgment and rapid learning (Lee and Edmondson, 2017). Protect those arenas from premature escalation. Reserve managerial authority of the Mark II variety, the structuring and sensemaking kind, for shaping interfaces and norms rather than micromanaging tasks (strategy+business, 2022).
Finally, resist the temptation to starve essential coordination functions in the name of purity. Valve’s avoidance of dedicated customer support and marketing was not costless. Specialized teams that carry institutional knowledge and manage external feedback loops are part of the spine. The point is not to rebuild layers for their own sake. It is to prevent avoidable bottlenecks where the cost of ambiguity is high (Foss and Klein, 2022).
The practical read for 2020s leaders
The debate over hierarchy versus network is not a morality play. It is a design choice under uncertainty. The definition of self-management is clear, and its promise is real, especially for complex, nonroutine work (Lee and Edmondson, 2017). The record of firms that went all-in on flatness shows recurring limits, from informal power filling the vacuum to performance challenges that force a reset (Foss and Klein, 2022; Groth, 2020). Managers still matter, albeit with authority that shifts from blanket control to targeted commitments and clear escalation rights (strategy+business, 2022).
Before you flatten, write down three things. First, which decisions need a named decider under time pressure, and how that person is chosen. Second, which functions must exist as durable teams because the cost of getting them wrong is high. Third, where you want the network to run without interruption because learning speed beats tidy lines. That spine will keep your network upright when the work gets messy, which is exactly when leadership is supposed to count.
Sources
- Lee and Edmondson (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing. Defines SMOs and argues hierarchy suits stable execution while self-management can aid complex problem solving. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059
- Foss and Klein (2022). Who's the boss? The persistence of entrepreneurial hierarchy in flat organizations, Journal of Organization Design. Shows flat firms often retain informal hierarchy and struggle in dynamic settings, with the Valve case as illustration. https://link.springer.com/article/10.1007/s41469-022-00132-8
- Bernstein et al. (2016). Beyond the Holacracy Hype, Harvard Business Review. Describes the polarized debate and notes Zappos as the largest holacracy implementation at the time. https://hbr.org/2016/07/beyond-the-holacracy-hype
- Groth (2020). Zappos has quietly backed away from holacracy, Quartz. Reports the 2015 buyout where 18% left and the 2017 shift to an internal market with managers reintroduced. https://qz.com/work/1776841/zappos-has-quietly-backed-away-from-holacracy
- strategy+business (2022). Can bossless management work? Summarizes Foss and Klein’s view that near-bossless firms are not generally superior and distinguishes two types of managerial authority. https://www.strategy-business.com/article/Can-bossless-management-work