I ran a session once with a senior leadership team that had spent six months developing a new values framework. They had invested real time in it. The process had been thorough: workshops with the broader leadership population, consultation with employees across levels, careful drafting and redrafting. The final values were good. They described something genuinely aspirational that most of the people in the room actually wanted to be true of their organisation. The team was proud of the work and invested in making it real.
About forty minutes into the facilitation, something happened. The most senior leader in the room, the CEO, was challenged on a recent decision that appeared to be in direct tension with one of the values the team had just spent months developing together. The challenge was made carefully, by someone with real standing in the room, in language that was respectful and specific. The response was not engagement. It was, delivered in a tone that ended the conversation, a brief explanation of why the decision had been necessary under the circumstances, followed by a decisive redirect to the next agenda item.
The rest of the team registered it. Nobody said anything. The session continued. The values work continued after the session. But the values had, in that moment, been defined more clearly than any workshop could define them. What the organisation now knew, in the part of itself that reads these signals accurately and continuously, was that the values applied to everyone except the person who had the most power and who therefore could not be challenged on them. Which means, in practice, they applied to no one. The gap between the aspiration on the wall and the behaviour in the room had been made visible in two minutes of unscripted interaction in a way that would take years of cultural work to close.
What Edgar Schein actually said
Edgar Schein’s formulation is worth stating with precision, because it is frequently quoted in ways that lose its sharpness. Culture is not primarily established by what leaders say about values. It is established by what they pay attention to, measure, and react to. By what they reward and what they tolerate. By how they behave in moments of difficulty, pressure, and ambiguity, when the social performance of leadership and the actual exercise of it most clearly diverge.
This formulation has a specific and uncomfortable implication. It means that the leader who says they value psychological safety while visibly dismissing challenges to their thinking in team meetings is not merely being inconsistent. They are actively building a culture in which psychological safety is recognised as aspirational language rather than operational reality, and in which the people around them are learning, correctly, that the invitation to challenge is conditional on the challenge being welcome. The learning is not happening because the leader intended to teach it. It is happening because the organisation is reading the behavioural evidence with considerable more accuracy than it is reading the communication.
The concept of the leadership shadow captures the specific mechanism through which this happens. The shadow is not the leader’s deliberately constructed influence on the organisation. It is the pattern of influence they exert simply through being who they are and behaving as they do, independent of any deliberate communication or culture-building effort. The shadow is not optional. It cannot be switched off for inconvenient moments. A leader who believes they are sending no signal about a particular dimension of the culture is, in fact, sending a very clear signal: that dimension does not warrant their attention, which is itself a highly consequential cultural signal.
The specific mechanisms through which shadow travels
Understanding why leader behaviour has such disproportionate organisational influence requires attending to the specific mechanisms through which that influence propagates through organisational systems.
The first is imitation. Albert Bandura’s foundational work on social learning theory established that humans acquire a significant proportion of their behavioural repertoire through observation and modelling of high-status individuals. The most senior person in any organisational context is the highest-status model available. Their behaviour, particularly their behaviour under pressure and in novel or difficult situations, is being observed and metabolised by everyone in the organisation who is close enough to observe it directly. The key word in Bandura’s framework is modelling: people do not simply note what the high-status person does. They develop an internal model of how this kind of person operates in this kind of situation, and they deploy that model when they encounter analogous situations themselves. This is why leadership behaviour at the top tends to reproduce itself throughout the organisation with remarkable fidelity, even when the leaders at lower levels have not been explicitly told to follow the senior leader’s approach.
The second mechanism is incentive signal. What leaders pay attention to communicates what is important. What they ignore communicates what can safely be treated as unimportant. What they reward directly, through formal recognition, through promotion decisions, through the allocation of high-visibility opportunities, communicates what kinds of behaviour the organisation is actually trying to produce. And crucially, what they tolerate communicates the real floor of acceptable performance rather than the stated floor. The organisation that formally espouses a culture of innovation but consistently promotes people who have safe execution records rather than people who have taken intelligent risks and sometimes failed is communicating, through the incentive signal, that the culture of innovation is an aspiration that the actual incentive architecture does not support.
The third mechanism is what might be called permission structure. When a senior leader behaves in a particular way, they are establishing, whether they intend to or not, the social permission for that behaviour throughout the organisation. If the most senior leader in a meeting interrupts people, others receive implicit permission to interrupt. If the most senior leader responds to bad news with anger or blame, others learn that honest early reporting is less safe than managed disclosure. If the most senior leader avoids difficult conversations and manages around conflict rather than through it, others understand that this is the operating norm regardless of what the culture documents say about direct and honest communication.
The amplification problem and why leaders consistently underestimate it
Senior leaders routinely and dramatically underestimate how much their behaviour is amplified by the organisation. A minor impatience that the leader experiences as a passing mood, barely worth their own attention, is experienced by those around them as a signal about what kinds of input are welcome and what kinds are not. A joke at a colleague’s expense that the leader forgets immediately is remembered by the target, by everyone who witnessed it, and often by a wider circle who hears about it afterward. An off-hand comment about a colleague’s performance becomes the basis for how that colleague is perceived by everyone who heard the comment, modified by their own relationship with the person being discussed and by how well they know the leader’s tendency toward casual hyperbole.
This amplification is not the product of malice or organisational dysfunction. It is the rational response of a system that needs to read its most powerful members accurately in order to navigate effectively. People in organisations are engaged in a continuous process of gathering information about what is valued, what is safe, and how to succeed in this particular environment. Senior leaders are the highest-signal sources of this information available to them. Their behaviour is therefore observed with much more attention and interpreted with much more significance than the leaders themselves typically recognise when they are producing it.
The practical implication is uncomfortable. Leaders cannot take comfort in the distinction between the behaviour they intend and the behaviour the organisation observes. The organisation is observing the behaviour, not the intention. The leader who intends to be collaborative but whose actual behaviour in high-stakes conversations consistently follows a directive pattern is building a directive culture, regardless of their stated intentions and regardless of the sophistication of the culture-building programmes their organisation runs.
The specific behaviours that carry the most cultural weight
Not all leader behaviours carry equal weight in shaping organisational culture. The ones that carry the most weight are, reliably, the behaviours that occur at the boundaries between the leader’s stated values and their actual interests, particularly under conditions of pressure, uncertainty, or threat.
How the leader responds when given information they did not want to hear. This single behavioural pattern does more to determine whether an organisation has functional early warning systems than any amount of speak-up culture work. The leader who receives unwelcome information with genuine curiosity, who protects the person who delivered it and visibly values the honesty, is building an information environment that will surface problems early. The leader who makes the bearer of bad news regret bringing it, through any combination of anger, dismissal, or subtle marginalisation, is building an information environment that will filter bad news upward until it can no longer be filtered, which is to say until the problem has become a crisis.
How the leader handles the tension between moving quickly and including others. The leader who consistently shortcuts consultation when the timeline is tight is teaching the organisation that inclusion is a fair-weather value, available when convenient and dispensable under pressure. This is the precise condition under which the cultural message is most clearly transmitted: not when things are easy but when there is a genuine cost to maintaining the value.
How the leader responds to their own mistakes. The leader who acknowledges error directly and specifically, without deflecting into explanation or attribution, models for the entire organisation the relationship between honesty and safety. The one who consistently attributes their own failures to circumstances or to others’ shortcomings models a different and more corrosive norm: that accountability is asymmetric, applying to others but not to those with the most power.
What this means for culture change
The most common reason culture change programmes fail is not poor design of the programme. It is the absence of any serious attention to what the senior team is actually modelling. The gap between the culture the programme is designed to build and the culture the senior team is daily producing through their own behaviour is the primary source of the cynicism that accumulates around culture change efforts, and the primary mechanism by which the old culture reasserts itself regardless of the sophistication of the intervention designed to replace it.
Genuine culture change requires, as its first condition, that the senior team engage honestly with the gap between the culture they are describing and the culture they are producing through their own daily behaviour. This is not a comfortable engagement. It requires a quality of honest self-assessment that most leadership teams have not built the relational conditions to support, and a quality of skilled external facilitation that can hold the team in that self-assessment without the conversation collapsing into managed affirmation or defensive justification.
But it is the necessary starting point. The culture change that does not begin here is working on the downstream symptoms while leaving the upstream source untouched. The leadership shadow will continue to define the culture regardless of the sophistication of the intervention that is designed to change it, because the shadow operates through mechanisms that are both more continuous and more powerful than any designed programme. Changing the culture means changing the shadow. And changing the shadow means changing the specific behaviours that cast it, in the specific moments when those behaviours are most consequential and most difficult to change.
Culture is not what is written on the wall. It is what happens in the room when the most senior person walks in, and what changes when they leave.