Why Silos Survive Every Restructure

Every organisation I have worked with has a silo problem. The language varies — "alignment challenges", "cross-functional friction", "communication breakdowns" — but the underlying dynamic is the same. Units that should be operating as a single system are instead operating as separate ones, each optimising for its own priorities, each carrying its own picture of what is happening and what matters.

The standard response is structural. Reorganise. Create cross-functional teams. Install coordination mechanisms. Mandate collaboration. Hire someone to run integration.

And then, a year or two later, the silos are back. Sometimes in the same form. Sometimes in a slightly different configuration. But back.

Silos are not a structural problem. They are an inter-subjective reality problem. Until leaders understand that distinction, every restructure is just rearranging the furniture.

What silos actually are

A silo is not primarily defined by its org chart position or its budget boundary. It is defined by its story.

Every unit in an organisation — every function, business line, geography or team — develops, over time, its own shared understanding of what matters, what the threats are, who the allies and adversaries are, and what kind of behaviour is rewarded and punished. This is what I have been calling inter-subjective reality in this series: the shared world that exists between people, constructed through accumulated experience, language and meaning-making.

When that inter-subjective reality becomes sufficiently distinct from the one operating in the unit next door, you have a silo. Not because someone drew a box on an org chart, but because two groups of people are now effectively living in different worlds — with different stories about the same events, different interpretations of the same data, and different instincts about what a crisis means and what it requires.

This is why restructuring does not fix silos. You can move people around, redesign reporting lines, create new integrating roles and governance forums — and within months the new structure will have generated new silos, because the underlying dynamic that produces them has not been addressed. The structure is objective reality. The silo is inter-subjective. Changing one does not automatically change the other.

The crisis management problem

Nowhere is this more consequential than in crisis.

Traditional crisis management structures — Gold, Silver and Bronze command in the UK public sector model; Incident Command Systems in US emergency management; functional crisis teams in corporate environments — are built on a set of assumptions that made sense in a more predictable world. The hierarchy provides clarity of authority. The tiered structure ensures decisions are made at the appropriate level. The process defines communication flows and escalation paths. In a relatively stable environment, facing disruptions of known types, these structures work.

But in the kind of complex, rapidly evolving environment described in the previous article — where the disruption does not respect the categories the structure was designed around, and where the pace of change exceeds the speed of the hierarchy — these structures create as many problems as they solve.

The information problem

In a hierarchical crisis structure, information flows upward and decisions flow downward. This is designed to ensure that the people with the most authority have the most complete picture. In practice, under pressure, the opposite tends to happen. Information gets filtered at each level — consciously or not — by people who are managing their own subjective reality: their fear of delivering bad news, their identity investment in their unit's performance, their assessment of what the person above them wants to hear.

By the time a picture of reality reaches the Gold commander or the executive crisis team, it has often passed through enough subjective filters to bear a limited relationship to what is actually happening on the ground. The leader is making decisions based on a version of objective reality that has been shaped, at every level, by the subjective realities of the people who carried it upward.

The silo problem under pressure

Crisis also activates silo dynamics in their most acute form. When pressure increases, units default to self-protection. Information that might reflect badly on one function does not flow naturally to another. Decisions that touch on another unit's territory generate friction and delay. The coordination mechanisms that work adequately in normal conditions seize up precisely when they are most needed.

The result is what organisation theorists call "sensemaking failure" — the collective inability to construct an accurate shared picture of what is happening — at exactly the moment when a shared picture is most critical. Organisational units are, in effect, operating in their own subjective realities, without the inter-subjective reality — the shared story — that would allow them to function as a single system.

Team of Teams — and what McChrystal actually understood

When General Stanley McChrystal describes what the Joint Special Operations Command had to learn to defeat Al-Qaeda in Iraq, he is describing, in operational terms, exactly this problem.

The enemy was not more capable. It was more adaptive. It operated as a network — fluid, decentralised, fast — while JSOC operated as a hierarchy — structured, coordinated, but slow. The hierarchy could execute a plan brilliantly. It could not adapt to a situation that changed faster than the planning cycle.

McChrystal's response was not to abandon structure. It was to fundamentally change the nature of the shared reality operating within it.

Traditional Crisis Structure vs Team of Teams

Dimension	Traditional Crisis Structure	Team of Teams
Information flow	Upward through hierarchy; filtered at each level	Lateral and continuous; shared across the network in real time
Decision authority	Concentrated at the top; escalation required for most decisions	Pushed to the edge; leaders empowered to act within shared intent
Shared picture	Each unit holds a partial view; synthesis happens at the top — slowly	Shared consciousness maintained across units; everyone sees the same picture
Response to novelty	Escalate to a level with authority; wait for decision; execute	Those closest to the problem act immediately within shared understanding
Silo dynamic	Reinforced under pressure; units default to self-protection	Broken by design; cross-team relationships built deliberately and continuously
Inter-subjective reality	Fragmented; each unit has its own story about the crisis	Unified; a single shared story about what is happening and what it requires
Leader's primary role	Command, process ownership, decision-making	Steward of shared consciousness; empowerer of edge decisions

The operational insight — that shared consciousness enables empowered execution — is the reason Team of Teams works in complexity. But there is a deeper truth beneath it that McChrystal identifies and that is rarely foregrounded in how organisations apply his ideas.

Shared consciousness is not created by sharing information. It is created by sharing meaning. And meaning — the inter-subjective reality of what this situation is, what it requires of us, and who we are in it — can only be created in conditions where people can tell the truth.

Psychological safety — the enabling condition, not the cultural aspiration

Amy Edmondson's research on psychological safety — the belief that one can speak up, challenge, disagree and share bad news without fear of punishment or humiliation — is usually presented as a leadership virtue. A quality that good leaders cultivate because it makes their teams feel better and perform more consistently.

This framing, while accurate, significantly undersells what is actually at stake. In the context of this argument, psychological safety is not primarily a cultural aspiration. It is the enabling condition for accurate shared reality.

Challenging subjective reality

Every leader carries a subjective reality — a set of beliefs, assumptions and interpretations through which they make sense of events. Under pressure, that subjective reality is most likely to distort. The situation most threatening to the ego is the one most likely to trigger defensive processing — filtering information to confirm existing beliefs, dismissing challenge as disloyalty, mistaking certainty for clarity.

Psychological safety is the mechanism by which the leader's subjective reality can be challenged before it becomes the organisation's official version of events. It is the condition under which someone in the room can say: "I think we're reading this wrong" — and have that statement received as valuable rather than punished as insubordinate.

Without it, the leader's subjective reality — their filtered, ego-shaped interpretation of objective events — calcifies into the shared story. The team adopts it not because they believe it but because the environment makes disagreement too costly. The result is an inter-subjective reality built on a single person's unexamined inner world, distributed across an organisation that has collectively agreed to stop questioning it.

This is one of the most consequential failure modes in leadership. And it is entirely invisible to the leader at the centre of it, because the absence of challenge feels like alignment.

Challenging inter-subjective reality

The same dynamic applies to the shared story itself. Inter-subjective reality — the collective beliefs a team holds about who they are and what is happening — can become as distorted as any individual's subjective reality.

Once a shared story hardens, it stops being a picture the team consults and becomes the lens through which every new piece of information is interpreted — evidence that fits is absorbed, evidence that contradicts is explained away. This is groupthink in its organisational form: not an absence of intelligence, but a collective agreement, usually unspoken, about what will and will not be seen. Psychological safety is what keeps the shared story open to revision — the condition under which a team can say "the story we have been telling ourselves no longer matches what is in front of us" before the gap between the two becomes a crisis.

The leader's role — command, process, decisions?

Which brings us to the question underneath all of this: if the shared reality matters more than the structure, what exactly is the leader for? The instinct is to answer in the familiar terms — command, process, decisions. Each is worth taking in turn, because none should be discarded, and all three need reframing.

On command.

Command is not the problem. There are moments when a team needs a clear direction held by one person, and the absence of it is its own failure. The leader who abdicates command in the name of empowerment — who retreats into facilitation when what is needed is decision — is not serving their team. They are leaving them exposed.

But command is only as good as the information that informs it. A leader who commands clearly, based on a picture of reality that has been filtered by hierarchy and distorted by their own subjective defensiveness, is exercising authority confidently in the wrong direction. The quality of the command is not a function of its clarity. It is a function of the accuracy of the reality it is based on.

This is why the leader's relationship to their own ego — their capacity to remain genuinely open to challenge rather than performing openness while actually closing down — is not a soft leadership quality. It is a determinant of whether their command capability is an asset or a liability.

On process.

Process is the organisation's memory made operational. A good crisis management process encodes the lessons of previous disruptions, reduces the cognitive load on people operating under pressure, and creates shared coordination without requiring everything to be decided in the moment.

The problem is that process, like hierarchy, was designed for disruptions of known types. It is a crystallised response to a previous version of the environment. In a genuinely novel situation — the kind of situation that defines the current international environment — process provides a scaffold, but it cannot substitute for judgement. And judgement, under pressure, requires the same conditions as command: accurate shared reality, honest information flow, and a leadership team that can challenge its own assumptions rather than executing its own pre-existing story.

The leader's role in relation to process is therefore not to own it but to understand when it is helping and when it is constraining. When the process is creating coordination and reducing cognitive load: use it. When it is creating false confidence that a novel situation is being handled by a familiar method: override it. That distinction requires the clarity of self-awareness that we have been discussing throughout this series.

On decisions.

McChrystal's formulation is the most useful I have encountered: "Eyes on, hands off." The leader's job is not to make every decision. It is to maintain a clear enough picture of the whole — and to have invested enough in the relationships and shared understanding of the people around them — to know which decisions to make and which to empower others to make.

The leader who makes every decision is not exercising leadership. They are creating a bottleneck. Every decision that travels to the top for resolution is a decision that did not get made by the person closest to the information — and a moment of delay in a system that is already under pressure.

But the leader who delegates without investing in shared understanding is not empowering their team. They are abandoning them. Empowered execution in a Team of Teams structure only works if the people at the edge genuinely understand the intent clearly enough to make decisions consistent with it. That shared understanding — the inter-subjective reality of what we are doing and why — is the leader's primary product. Not the decisions themselves.

The leader's role in a complex environment is to be the trustworthy steward of the shared reality. To hold a picture of the whole that is accurate enough to inform good decisions, honest enough to surface uncomfortable truths, and stable enough to hold the team's identity together when the environment is at its most threatening.

This requires all three things this series has been exploring: the self-awareness to manage one's own subjective reality rather than be driven by it; the storytelling capacity to build and maintain the inter-subjective reality of the team; and the psychological groundedness to create the conditions under which both can be genuinely challenged.

Command, process and decision-making are the instruments of leadership. But they are only as good as the inner quality that wields them. And that inner quality — the ego examined, the identity stable, the shared reality held with honesty and care — is the thing that almost no leadership development programme addresses, and that every genuinely effective leader I have known has worked, in their own way, to develop.

A final thought

Silos will not be solved by restructuring. Crisis management will not be improved only by better process. And organisational resilience will not be built by plans alone.

What breaks silos is a shared inter-subjective reality strong enough to hold across unit boundaries — built deliberately, maintained consciously, and protected by the psychological safety that allows it to be challenged before it calcifies into comfortable fiction.

What makes a crisis structure effective in genuine complexity is not its hierarchical clarity but its capacity for honest shared sensemaking — the ability to construct and continuously revise an accurate picture of what is happening, fast enough to act on it, with enough edge empowerment to respond without everything travelling to the top.

And what makes a leader effective in that environment is not primarily command authority, process mastery or decision speed. It is the quality of the reality they hold — and the depth of the inner work that allows them to hold it honestly.

That is the argument this series keeps returning to, from different angles. Not because it is comfortable. But because, in thirty-five years of leading and observing leaders in some of the most demanding environments in the world, it is the one that keeps proving true.

Why Silos Survive Every Restructure — And What Actually Breaks Them.