What structure really is, and why it matters

What we mean by "structure"

Most people who read this newsletter have encountered the word "structure" used in ways that go beyond the organizational chart. A reporting relationship is a structure. So is a misalignment between accountability and authority. So is an incentive program that rewards individual performance while expecting collective behavior. So, sometimes, is a cultural norm that makes certain conversations undiscussable regardless of who is in the room.

That breadth raises a legitimate question: what actually qualifies as a structure, and what doesn't? Without a working definition, the word risks becoming a placeholder for anything that is hard to change or has a predictable impact on behavior — which is too loose to be analytically useful.

I am using this term in a specific way that may or may not align with how it is used in other conceptual models, so a brief definition is in order.

It is worth noting that a structure is not necessarily something tangible. What makes something a structure is not its physical nature but its effect: it shapes behavior predictably, recurrently, and independently of who is involved. Some structures are visible and formally designed. Others are made up of shared assumptions, informal expectations, and the accumulated weight of how things have always been done. Both qualify.

In other words, the question is not "what kind of thing is this?" but "does this thing function as a structure?" If it shapes behavior predictably, recurrently, and independently of who is involved, it qualifies. If it doesn't meet all three conditions, it belongs in a different analytical category.

How structure and system relate

Every structure exists within a web of systems — overlapping, nested, and more elusive and often fleeting than the structure itself.

Consider a single team in an organization. The team itself is a system. But so is every pair of relationships within it — two colleagues who habitually align before meetings, two others who consistently talk past each other. So is every coalition that forms around a project, every recurring dynamic between the team and its manager, every pattern that emerges when the team is under pressure versus when it is not. The number of systems operating within and around a single team is not fixed; it is potentially limitless because any set of interacting components that produces recurring outcomes qualifies.

The structures within that same team are far more bounded. Not every interaction pattern qualifies — only those that shape behavior predictably, recurrently, and independently of who is involved. That is a more selective test, and it produces a shorter and more specific list. The team may have dozens of overlapping systems operating simultaneously. Its structures are fewer, more identifiable, and more directly actionable.

That contrast is what makes system and structure different in kind rather than just in scale. System is generative — it emerges wherever components interact. Structure is selective — it is identified by what it does to behavior.

That contrast also has an important practical consequence. When you intervene in one structure, you are intervening in one node of a potentially infinite web of interacting systems. The rest of that web continues to operate, which is part of why structures can reconstitute themselves even after being deliberately dismantled. We will return to that point.

The relationship between structure and system is precise: every structure is a system, but not every system is a structure. All ice is water, but not all water is ice. A system becomes a structure when its patterns are stable and persistent enough to shape behavior predictably, regardless of the intentions of the individuals operating within it. Predictably, but not deterministically. A structure does not compel behavior; it makes certain behaviors significantly more likely than others. Someone who resists will not necessarily conform, but they will have to struggle against the inertia of the structure and the cumulative pressure it exerts toward conformity. That struggle has a cost, and most people, most of the time, do not sustain it indefinitely or find themselves pushed out, one way or another.

In sum, structure is what emerges from systems when their patterns have hardened into conditions that people have to navigate whether they choose to or not.

Is a structure good or bad?

A structure is neither good nor bad in itself. It simply is. The evaluative question is always relational — good or bad relative to what, and for whom.

The first dimension of evaluation is purpose

A structure that reliably produces what it was designed to produce is, by that measure, a well-designed structure. But this raises a prior question that is easy to overlook: whose purpose? In practice, organizational purpose is not a neutral or collectively determined thing. It is defined by those who hold the power to define it — senior leaders, owners, investors. A structure that fits their purpose well is not automatically one that fits the purpose of everyone else operating within it, or everyone affected by it. The design question and the interests question are not the same question, even when they appear to be.

The second dimension of evaluation is ethical

What is the structure ultimately contributing to? That question cannot be answered by pointing to intent. A structure that serves its designers' purposes well can still be part of a chain of effects that produces harm elsewhere: to the people operating within it, to those who use the organization's products or services, to communities, to broader social or environmental conditions. The structure may not be the only cause, and it may not be the proximate one. But it is part of what is producing the outcome.

Why this matters in practice

The most practically important implication of the relationship between structure and system is what it tells us about structural change.

Dismantling a structure — changing role definitions, realigning accountability and authority, redesigning incentive programs — does not automatically dismantle the systems it was embedded in. The components still exist. The interactions continue. The habits, informal relationships, and power dynamics that developed under the old structure don't disappear because the formal design changed. If those interactions continue to produce recognizable and recurring patterns, the systems are still operating — often reproducing something that closely resembles the structure that was supposedly dismantled.

This explains one of the most common and frustrating experiences in organizational life: we changed the structure, but nothing really changed. That outcome is not a failure of implementation or a failure of will. It is a predictable consequence of changing the structure without adequately accounting for the systems it sits within. Structural redesign that does not attend to the systems it is embedded in is likely to disappoint — not because the redesign was wrong, but because it was incomplete. But the practical stakes go further than examining whether a structure works as intended. They also require asking what it is working toward, and at whose expense.

Alexander Spradling, PhD, is an organizational consultant and executive coach based in Austin, Texas. He writes Why Organizations Misbehave.