It's Time to Embrace Hierarchy

How to Save Self-Management from Itself

Most writing about self-management focuses on why it beats hierarchy: top-down command stifles creativity, disempowers people, and can’t keep up with complexity, so we distribute authority, flatten the org chart, and trust people to manage themselves. There’s truth in that.

But it carries a serious limitation the movement rarely names—it thinks about governance primarily in contrast to hierarchy. Holacracy, sociocracy, Teal, etc.—the posture is the same: we’re building something that isn’t hierarchy.

But I think this negative definition constrains what self-management can (and should) become.

So here’s the argument: self-management needs something bigger than itself—a general theory of governance that accounts not just for the self-managed mode, but for all the ways humans organize work.

It needs that theory precisely because real organizations are never just one kind of system.

Organizations Are Nested

When people describe an organization’s governance, they reach for a single label: “we’re hierarchical,” “we’re flat,” “we’re matrix.” But any organization of meaningful size is a composite—smaller organizations nested inside each other, each with its own logic and its own implicit style of governance.

• Accounting might run as a tightly controlled, rule-driven operation—what Larry Constantine would call a Closed system.

• The creative team thrives as a Random system, on individual initiative and autonomy.

• Product engineering runs Open, on collaboration and negotiation.

• A small, long-tenured team coordinates in a Synchronous mode, on deep shared assumptions.

And this nesting is inevitable. Any organization of meaningful size develops subgroups—departments, teams, working pairs—because people can only coordinate closely with so many others at once.

Each subgroup has the same features that make its parent an organization: a membership boundary, coordination patterns, norms, decision-making.

So each is itself an organization, and the nesting runs all the way down to individuals and all the way up through industry, regulation, and society. It’s fractal.¹

The implication is the one most frameworks miss: if nested subgroups develop different governance logics, then prescribing a single mode for the whole organization works against the grain of how organizations actually function.

That’s a structural argument against governance monocultures of every kind—including the self-management monoculture.

The Self-Management Bias

Self-management has its own version of this blind spot. It assumes the distributed, participatory mode is universally preferable. Hierarchy is the problem—full participation is the answer. Always.

• But what happens when part of your organization genuinely needs tight procedural control for compliance?

• When a small team does its best work precisely because it’s been left alone?

• When a highly aligned team executes with extraordinary efficiency on shared assumptions?

The self-management framework, in its zeal to displace hierarchy, has no good language for saying: “Actually, this part of the organization should be governed differently, and that’s fine.”

The irony is painful to watch. In escaping the limits of one governance mode, self-management advocates end up imposing a different single mode on the whole organization.

They’ve replaced one form of homogeneity with another. The content has changed, but the structural mistake is identical: treating the organization as though it needs one unified operational logic from end to end.

Conclusion

The self-management movement surfaced the real costs of uncritical hierarchy. But the next step isn’t more self-management—and it isn’t a finished theory to replace it, either.

There’s no one best way to organize, and no one best way to analyze how we organize. Both targets move, because we do. What we get are steps. Some to the side. Some backwards.

What we need most are steps forward. And embracing hierarchy (i.e., reintegrating it) is one way to do that.

1. There is some good history behind this idea already.

• The subcultures literature—Joanne Martin, John Van Maanen, and others—has documented for decades that organizations don’t have single cultures but constellations of subcultures that overlap and compete.

• Lawrence and Lorsch showed in the 1960s that different parts of an organization face different environments and develop different internal structures—and that the core challenge is integrating across those differences.

• Mintzberg’s structural configurations describe the same reality from another angle, and research on multi-team systems studies how teams-of-teams coordinate when each has its own goals and dynamics.