Sometimes the leader is too big for the room.

Not physically, but in the way their presence affects everything in the space: how people speak, what they dare to say, which decisions suddenly feel different the moment they are present. In that situation, you can talk endlessly about behaviour, naming and discussing every individual statement and every individual moment of arbitrariness. Or you can do something fundamentally different: you can change the playing field itself.

Small system interventions sound harmless, but they touch the core.

They reduce arbitrariness not through confrontation, with all the resistance and escalation confrontation brings, but through form. You make decisions more transparent, so they no longer depend on whoever happens to hold the most influence at that moment. You make criteria more explicit, so discussions about what counts as a good decision no longer remain vague. You make roles clearer, so nobody can claim something falls within their mandate when it factually does not. And with that, impulse becomes more costly: it takes more effort to act outside the agreed structure than to move within it.

It is almost mathematical in its simplicity. In a room without fixed agreements, a strong personality can shape reality to their own preference, simply because nothing offers resistance. In a room with rhythm and record-keeping, that same personality suddenly has to move along rails. They can still push, that remains possible, but the train stays on the track. The impact of their pushing is bounded by a structure that does not depend on them.

Psychodynamically, this is what theory calls containment: a system’s capacity to hold tension without that tension having to occupy people, without it landing somewhere else, in bodies, in sleepless nights, in constant gossip that spreads because there is no formal place to name the tension. Rather than fear spreading through informal channels, tension can exist in an agenda, in a decision list, in a rhythm that is recognisable and predictable for everyone.

Research into organisational architecture supports why this approach is so effective, often more effective than direct behavioural interventions. Structural adjustments, such as clear decision-making processes and transparent criteria, prove more durable in their effect than one-off conversations about behaviour, because they do not depend on individuals’ ongoing effort to confront the issue. A structure keeps working, even on the days nobody has the courage to say anything.

Think of a simple decision table with clear questions asked the same way every time, regardless of who is sitting at the table. Think of one place where decisions are recorded, accessible to everyone who needs to work with them. Think of a dashboard that belongs not to one person but to the task itself, so nobody can claim it as their personal instrument. Think of an explicit escalation route: when does something go to which table, and why, so nobody can claim they did not know where a decision was supposed to be made.

The pitfall is real: the leader can hijack the instruments.

This is a warning that should not be skipped, because it makes the difference between an intervention that works and one that backfires. A dashboard can become a weapon in the hands of someone who controls it without challenge. A decision list can be filled selectively, so it creates the appearance of transparency while in reality being just as manipulable as the old, informal system. That is why shared ownership is crucial to the success of any system intervention. Not one administrator with the power to make changes without accountability, but shared management in which multiple people have access and a say. Not one source of truth that disappears into a drawer or is only accessible to whoever holds the passwords, but a source that is visible and consultable for everyone entitled to it.

It is wise to name this pitfall explicitly when setting up the intervention, not only once it has already gone wrong. Ask yourself, and possibly others, in advance: who can adjust this structure in future, and under what conditions? If the answer is that one person can do so unilaterally, you have effectively recreated the old problem, only now wrapped in a new instrument with the appearance of objectivity.

Do not make it personal when you introduce this intervention, even though the occasion for it is indeed personally related to the behaviour you are trying to limit. Your story is not: I don’t trust you, so we are building this structure. That story immediately invites a defensive reaction and frames the intervention as an attack. Your story is: we want to be able to execute faster and have fewer misunderstandings, so we are standardising. That is a story that sounds beneficial to everyone, including the leader themselves, because nobody can officially be against speed and fewer misunderstandings.

Choose one small form today that fits your context. Not everything at once, because an overly ambitious attempt to restructure the entire culture in one go provokes more resistance and is harder to sustain than a small, concrete start. One thing that removes arbitrariness from the air, one structure that makes an immediately noticeable difference without being experienced as a major upheaval.

And then watch what happens, not only in processes and in any figures you might track, but in people, in how they behave and feel. You will often notice shoulders dropping slightly, a physically recognisable signal of relief. Not because the problem is entirely gone, since that is often not yet the case after one small intervention. But because there is structure again, a place people can rely on, independent of the mood of whoever happens to be in charge at that moment.

And ask yourself afterwards: which small structural agreement would remove the most arbitrariness for you, with the least resistance and the most immediate effect?

The answer to that question is usually simpler than you expect, and that very simplicity is what makes the intervention achievable in an environment where large, dramatic changes often get stuck on the resistance they provoke themselves.

Notes for those who wish to read further:

1. Wilfred Bion, Learning from Experience (1962, Heinemann). The psychoanalytic origin of the concept of containment, and how structure can hold tension without it causing harm elsewhere.
2. Jay Galbraith, Designing Complex Organizations (1973, Addison-Wesley). On how organisational structure steers behaviour, often more effectively than direct instructions or confrontations.
3. Donella Meadows, Thinking in Systems (2008, Chelsea Green). On how small, targeted interventions in a system can have larger and more lasting effects than large, direct interventions.
4. Gerard Endenburg, Sociocracy as Social Design (1998, Eburon). On decision-making structures that distribute power through form rather than personal authority.
5. Erik Hollnagel, Safety-II: The Past and Future of Safety Management (2014, Ashgate). On how systems build resilience through structure, rather than relying on individual vigilance.