The Gentlemen's-
Agreements Test.

The hardest question in the Five-Layer Diagnostic is not about coding, or LOS, or staffing. It is about politics. Specifically: can you name three informal agreements between your heads of department — or between your clinical and operational leadership — that prevent a specific operational improvement. The kind of agreement that would never appear in meeting minutes but that everyone in the room understands.

Most leaders, asked this question for the first time, cannot name three. Some cannot name one. The inability is the finding.

This post is about why the gentlemen’s-agreements layer matters diagnostically, what kinds of agreements are typically present in hospitals, and what it means — for you as a leader — to be able, or unable, to see them.

Every hospital runs two organisations.

The first is the one on the org chart. Geschäftsführer at the top. Medical Director, Pflegedirektor, and Leitung Controlling reporting into her. Chefärzte reporting into the Medical Director. Stationsleitungen into the Pflegedirektor. Boxes, lines, titles. This is the organisation that is published, printed in the annual report, displayed in the lobby. It is also — and this is the uncomfortable part — not the organisation that actually decides whether operational change sticks.

The second organisation is the one that runs on relationships, debts, shared history, and informal alliances. It doesn’t appear on the org chart. It doesn’t appear in meeting minutes. Nobody describes it in public — it is the kind of knowledge that seasoned staff have and that new CEOs learn by accident, usually by proposing a change that inexplicably fails.

In the informal organisation, the Chefarzt Chirurgie has more power than the Medical Director on matters related to OR scheduling, capital equipment, and surgical staff rotation — not because he formally does, but because he and the Leitung Controlling have a fifteen-year working relationship and neither will actively contest the other. The Senior Pflegekraft on Station 3 — Andrea — has more influence over how nursing policy actually plays out than the Pflegedirektor, because every ward looks at Andrea’s reaction before adopting anything. The Chefarzt Innere has formal parity with the other heads of department but is functionally isolated, because he stopped attending the informal Friday-evening dinner eight years ago after a specific incident that nobody will discuss.

The informal organisation is not good or bad. It is always present. The question is whether the formal leadership has learned to see it.

Three categories of agreement — with examples.

When the Five-Layer Diagnostic asks you to name three gentlemen’s agreements, it is not asking for a complete political ethnography of your hospital. It is asking whether you have sufficient visibility into the informal layer to name three specific agreements — by named participants — that are actively blocking a specific operational improvement. That is a much narrower question than it sounds, and the narrowness is the diagnostic.

Agreements typically fall into three categories.

• Protective agreements

  Two leaders have implicitly agreed not to challenge each other on a specific topic, usually because both have something to lose from the challenge. The Chefarzt Chirurgie does not challenge the Leitung Controlling’s allocation methodology because he has benefited from it historically; the Leitung Controlling does not challenge the surgical department’s capital requests because a past challenge produced political fallout nobody wants to repeat. The agreement is invisible until someone tries to introduce a new allocation methodology and both parties unite — without coordination — against it.

• Jurisdictional agreements

  Two leaders have divided a specific domain between them in a way that is not formalised but is operationally binding. The Medical Director does not involve herself in departmental scheduling decisions; in exchange, the Chefärzte do not contest her authority on cross-departmental clinical governance. The agreement is invisible until someone proposes standardised scheduling across departments and both parties experience the proposal as a threat, without being able to explain why.

• Silence agreements

  A specific topic is understood — by everyone in the leadership meeting — to be off-limits. A known failing of a specific senior leader. A recurring coding loss pattern that implicates a named physician. A compensation arrangement inherited from a predecessor that is structurally misaligned but that nobody will put on the agenda. The topic does not get raised, not because anyone is afraid, but because raising it requires political capital that nobody in the room has decided to spend. The agreement is invisible until a new CEO joins and asks, “why are we not talking about X?”

These categories are not exhaustive. There are also loyalty agreements, historical debts, family-of-origin patterns in founder-run clinics, and a specific category of agreements that exists in German hospitals around academic-clinical dual appointments that I won’t describe here because the examples would be too identifying. The point is that informal political agreements are structural to hospital life, they cover specific and diagnosable terrain, and they can be named if the leader is looking.

Why the test is specifically three.

The diagnostic asks for three. Not one, not five. The specificity is deliberate.

One is too few because almost every leader can identify one agreement after some reflection — often the most obvious one, frequently involving a figure who has already left the organisation. Naming one agreement proves only that the leader has some awareness that informal politics exist, which is necessary but not sufficient for diagnostic clarity.

Five or more is usually achievable only by leaders who have catalogued the informal layer systematically, typically because they have done so as part of a prior diagnostic or because they have been in the organisation long enough to have collected the data through scar tissue. Five names a deep diagnostic capability that most organisations don’t need.

Three is the diagnostically useful middle. It requires the leader to have moved beyond the one obvious example. It requires the leader to have examined multiple domains of the organisation — clinical, operational, financial, workforce. It requires the leader to have named specific agreements rather than gesturing vaguely at “hospital politics.” Three is sufficient to reveal whether the leader has genuine diagnostic visibility into the informal layer, and insufficient to be achievable by accident.

What the agreements actually cost — operationally.

The gentlemen’s agreements matter diagnostically because they are almost always the explanation for why improvement initiatives that technically make sense fail to implement. I have seen this pattern in every hospital I have worked in as a department lead, and in every consulting engagement since.

A hospital commits to reducing length-of-stay by 15%. The technical elements are straightforward — documented pathways, structured ward rounds, a 24-hour coding loop, weekend discharge authority. The commitment is real; the Geschäftsführer is behind it; the Medical Director is on board. The initiative launches. Six months in, most elements have stalled. LOS has moved 3%.

When you walk backwards from the stall, you find specific informal agreements that blocked specific elements. The weekend discharge authority ran into the Chefarzt Chirurgie’s implicit claim on Saturday OR scheduling, which created a cascade of staffing decisions that made weekend medical discharge impractical. The structured ward round required a named handover owner for each ward, which conflicted with an existing agreement between the Pflegedirektor and two Stationsleitungen about who owned handover. The 24-hour coding loop required Leitung Controlling to push coders into daily clinical contact, which he was reluctant to do because of a historical agreement with the Chefarzt Innere about the limits of controlling’s presence on wards.

Each of these agreements was invisible at the start of the initiative. Each was load-bearing in the sense that dismantling it required political capital the initiative’s leadership had not budgeted. The net result was that the initiative’s formal structure was compatible with the initiative’s goals, but the informal structure of the organisation was not — and the informal structure won.

What to do if you can name three — and what to do if you cannot.

There are two possible diagnostic situations, and they require different responses.

If you can name three agreements specifically, you have done the visibility work. The next question is not about seeing more — it is about deciding which agreements to leave alone and which to dismantle. Not every informal agreement is harmful. Some hold the organisation together in ways that formal structure cannot. The skill is in distinguishing load-bearing agreements (which you leave, because removing them causes cascading damage) from legacy agreements (which you examine, because they block improvement for no remaining reason). This distinction is the advanced work of the FLOW layer, and it is rarely taught in management training.

If you cannot name three, the first move is not to try harder at naming them abstractly. The first move is operational: walk the wards, attend the informal meetings, watch the coffee-break conversations. Informal agreements reveal themselves to leaders who spend time in the informal spaces of the organisation. They do not reveal themselves in formal review sessions, because formal review is precisely the context in which informal agreements are honoured by silence.

A leader who has not yet learned to see informal politics will try to fix this through process — more meetings, more surveys, more 360s. These don’t work, because the informal layer is specifically designed to be invisible to formal process. What works is presence, unscheduled time, and the willingness to ask the uncomfortable question: “what is the thing we are not discussing?”

The larger point — and why this matters for improvement.

The test is not really about the agreements themselves. It is about whether you, as a leader, have developed the diagnostic capacity to see your own organisation’s political constitution. Every improvement initiative — every CAPTURE fix, every FLOW redesign, every PRIME rhythm — will eventually run into this layer. Leaders who cannot see it will experience repeated, inexplicable failure. Leaders who can see it will make choices that account for the informal structure and consequently produce changes that hold.

The Schlüchtern transformation succeeded, in part, because the department lead spent the first eighteen months not trying to improve anything — spending that time instead understanding the informal constitution of the department and of the hospital around it. What agreements were load-bearing. Which were legacy. Which leaders owed each other what. Where the silent topics were. Only after that period did structural improvement begin. And because the improvement accounted for the informal structure, it held. The pattern is neither unique to that department nor unique to geriatrics; it is a general feature of hospital change work, documented formally in the Schlüchtern case [1,2] and observed across subsequent engagements.

Most consulting engagements fail because they skip this step. The engagement arrives, diagnoses the technical problems correctly, recommends the technically correct interventions, and then watches them stall — because the consultants never had access to the informal layer, and the client leadership could not see it either.

The Gentlemen’s-Agreements Test is the diagnostic shortcut. Three names. Specific examples. Out loud. If you can do it, you have the visibility to lead structural change. If you cannot, that is the first thing to fix — before any operational redesign is commissioned.