The Secretariat for Top Management vs. the One for the CEO

Peter Drucker, Management: Tasks, Responsibilities, Practices — Chapter 51: Top-Management Structure

Introduction

Say the word “secretariat,” and two pictures come to mind: a power bloc that steers the organization from behind the curtain, or a personal staff that manages the CEO’s calendar. But there is a third kind most people rarely picture — an organ that pushes top management to see new angles and ask new questions. Not something that feeds the operating body, but something that feeds the brain: the part of the company that looks at the future and the whole.

This organ is indispensable in theory, yet barely exists in practice. Getting its necessity recognized, its agenda set, and its work made durable all prove hard.

For nearly four years I led the management research center of a major Korean conglomerate, an organization meant to do exactly this. Our aim was to tell executives not what they wanted to know, but what they needed to know. We hunted for our own agenda and put it before top management. It was a low-odds fight: a question the executives had not asked rarely found a place. It felt like carving a masterpiece with care, then tossing it into a wide, still lake.

This chapter helped me understand why. The entire organization is built to nourish the operating body. No one is assigned to nourish the brain — and the few who try discover that the brain needs a different kind of food than anyone knows how to supply.

Top management’s work differs in kind from operations

Drucker insists top management is not a one-man job: the volume is too much for one person, and no temperament fits all the work demands. Henry Ford is his example. Ford grew under an effective two-man team, and the moment one man left and Ford held everything alone, the company began its long slide.

But the line that struck me hardest was different. Top management’s work differs from operating work not in degree, but in kind. Operations deal with the present; top management deals with the future and the whole. Operations perform within a business already defined; top management asks what it should be. The first task Drucker named earlier was exactly that — “What is our business, and what should it be?” — a question today’s operating metrics never surface.

If the work is different, the information that feeds it must be different too. Yet most organizations send the brain the same reports they send the body.

This is the secretariat’s dilemma. When top management treats operations as its own job, the staff beside it becomes a power: intervention borrowed from top management’s name turns into authority. Yet a staff that only relays information can never outdo the line, whose information is richer, better situated, and timelier. So the secretariat must serve top management’s own work — but the awareness that such work exists has to come first. That was my low-odds fight: in an organization that had yet to recognize top management’s work, a research unit pushed its agenda alone.

Three traps of the secretariat

How should a brain-feeding organ be set up? In Korean companies, three traps recur.

The first is power. One large Korean group ran a very strong head-office staff for decades, controlling the group’s portfolio, personnel, and investment. The problem was that it started making decisions on top of feeding information: an unofficial body sat above the affiliates’ boards and chief executives, deciding while they bore the consequences. The organ meant to feed the brain had begun acting as the brain. The model is not simply good or bad — it grew new businesses early, yet it weakened the affiliates’ autonomy and dulled their appetite for innovation. The brain’s work and the work of feeding it must stay separate.

The second is restraint. The opposite group keeps its central staff to a minimum, with the chief executive running the businesses directly. There is no abuse of power, but too little of the future-and-whole view reaches top management; the brain receives only the reports the line sends up. Drucker’s own case shows the cost. A maker of electrical equipment moved from steam turbines into jet engines and then atomic power, each organized as its own division. But to the utility buying them, these were not three businesses; each was simply another way to generate electricity. No one inside saw it, because every division’s staff served only its own division. By the time a unit created to study tomorrow’s market pointed this out, the company had lost much of its standing. Everyone was doing competent work, and still the company lost its market — not because it stopped, but while everyone was busy.

But the real difficulty is neither power nor size. The third is the agenda. A brain-feeding staff works in one of two ways. In the first, it finds its own agenda and brings it to the executives — telling them what they need to know. In the second, the executives set the agenda and the staff follows — answering what they want to know. My unit chose the first, and the odds were low. Drucker names the reason: genuine innovation never fits the existing organization structure. The agenda a staff should chase belongs to no operating unit, so through the logic of operations it is invisible and hard to legitimize. The second path makes the role clearer, but if the staff only answers questions the executive already holds, it can never offer a new perspective. A diet of only what one wants to eat produces malnutrition.

Here the core surfaces: what the secretariat needs is not authority, nor the power to decide, but the independence to find and wrestle with its own agenda. And that independence is decided by how it is measured. The moment a staff is judged by the same short-term yardstick as an operating unit, it can no longer do its work, because finding what one needs to know is not measured by the quarter. Under that yardstick, it supplies what people want to know instead. The brain starves not because no staff exists, but because the staff that does is pulled, to survive, into the logic of operations.

A brain that prepares before the fire

What, then, should the organ look like? Drucker’s prescription is concrete. Keep it small and held to the key issues. Admit only those who have proven they can perform in real work, and after five to eight years move them back into operating roles, so the staff never loses touch with reality. Strikingly, the strong head-office staff above matched this — it too drew proven talent and rotated it back — yet still turned into a power. The difference was one thing. What decides a secretariat’s fate is neither its people nor its size, but whether it makes decisions or asks questions.

A remark from the discussant who leads our management-research conversations comes closest to the chapter’s conclusion: anyone can respond after something breaks, and putting out fires can look like hard work — but real top management prepares before the fire starts. That is what feeding the brain means — seeing the change not yet on the income statement, and framing the question no one has yet framed.

Drucker closes in one line: the bottleneck is always at the head of the bottle. No company can do better than its top management allows. When the brain starves, even the healthiest body loses its way.

In closing

We tried to tell executives what they needed to know, but few organizations were ready to receive it. The problem was never the staff; the awareness had to come first — that top management is a distinct kind of work.

Picture your own company. What does top management actually receive — the same operating reports as the line, or the questions that ask about tomorrow? What is your company’s brain being fed — what executives want to know, or what they need to know? And has your organization ever given anyone the independence to find out?

From a Monday-morning Drucker reading group, 7 a.m.