When the CEO Becomes the Operating System

What looks like a delegation problem may actually be a decision-system problem.

Over the years, I have had versions of the same conversation with CEOs of growing companies. The details change, the industries are different, but the pattern is remarkably similar.

The CEO in this story is a composite of several of those conversations. I have changed circumstances and combined details to protect the people and companies involved, but the underlying problem is very real.

At some point, usually with more frustration than pride, the CEO says some version of the same thing:

“Everything has to go through me.”

The first few times I heard it, I thought I understood the problem. I assumed it was delegation.

I'm no longer so sure.

It Doesn't Look Like a Crisis

Imagine a CEO running a company that has been operating for more than a decade. Revenue has grown steadily, and what started as a small team has become an organization with experienced managers, defined responsibilities, processes, dashboards, and software supporting different parts of the operation.

From the outside, it looks like a company that has successfully moved beyond its founder-dependent stage.

The CEO experiences something different.

The larger the company becomes, the more questions somehow find their way back to him. A customer wants an exception. Sales wants to modify commercial terms. Two projects are competing for the same people. Finance isn't sure whether to approve something.

None of these issues is particularly dramatic. Most require five or ten minutes.

That is part of what makes the problem difficult to see.

There is rarely a single catastrophic bottleneck. Instead, there are dozens of small decisions scattered throughout the day, each requiring a little context that somebody believes the CEO has.

Eventually, his mornings begin before he gets out of bed. There are messages waiting from the previous evening, so he answers a few while having coffee and handles more throughout the day. Sometimes he opens the laptop again after dinner because somebody needs an answer before tomorrow morning.

I've heard CEOs describe vacations where nothing particularly serious happened, yet they still ended up taking calls from the hotel. There was no emergency that ruined the trip. The company simply continued requiring small pieces of them.

They can leave the office.

What becomes increasingly difficult is leaving the company.

I Used to Think This Was About Delegation

My initial reaction to these situations was predictable: the CEO needs to delegate more.

Sometimes that is exactly the problem. Some founders struggle to give up control. Others hire managers but continue making decisions for them.

But that explanation became less satisfying as I encountered companies with capable managers who genuinely had authority and still kept returning to the CEO.

One contradiction appears surprisingly often.

The CEO complains that managers bring too many decisions to him. Then he describes a situation in which a manager recently made a decision without consulting him.

Exactly what he says he wants.

And he stepped in.

So I started asking a different question:

“Was it a bad decision?”

Often, the answer is something like:

“Not exactly.”

The manager may have followed the company's normal logic. Based on the information available, the decision was reasonable.

The CEO simply knew things the manager didn't.

Perhaps he remembered an exception made for that customer months earlier. Maybe he knew the relationship had recently become fragile after a conversation that never made it into the CRM. He might know that another customer competing for the same resources was strategically more important than its current revenue suggested.

Individually, none of those details necessarily determines the answer. Together, they change how the situation looks.

That was when I began to wonder whether the problem was really delegation.

Delegating Authority Is Not the Same as Distributing Judgment

The decisions that repeatedly reach a CEO are often not the simple ones. Growing companies tend to become reasonably good at handling those through processes, policies, software, and delegated authority.

The difficult decisions live in gray areas.

Should we make an exception for this customer? Which commitment takes priority when two customers need the same resources? Should we accept a lower margin because an opportunity has strategic value? Should Sales promise something that Operations can technically deliver but probably shouldn't?

There is rarely a single rule that produces the answer. The decision depends on some combination of customer value, margin, history, risk, capacity, relationships, previous commitments, and experience.

A CEO who has spent ten or fifteen years building a company accumulates an enormous amount of this context without necessarily realizing it. He knows why certain rules exist, which customers are worth making exceptions for, which numbers need interpretation, and which teams are under more pressure than the dashboards suggest.

Much of that knowledge was never deliberately documented because, for most of the company's history, it didn't need to be.

People could simply ask him.

When there were fifteen employees, that might have been extremely efficient.

At fifty or a hundred, the same mechanism can become a constraint.

The Company May Already Have Plenty of Software

This is where I think the problem becomes easy to misdiagnose.

A company can have a CRM, accounting software, project-management platforms, dashboards, documented processes, and defined approval workflows and still depend heavily on a few people's judgment.

The issue is not necessarily missing information.

It may be the relationships between pieces of information.

A CRM can tell a manager how much revenue a customer represents, but it may not explain why an exception was granted nine months earlier. A project-management system can show that two projects need the same specialist, but it cannot necessarily tell Operations which commitment matters more. Finance can correctly enforce a policy without knowing when the reasoning behind that policy might justify an exception.

A company can become very good at capturing what is happening without becoming equally good at preserving how it decides what should happen next.

That reasoning often remains surprisingly informal.

Sometimes it remains inside a handful of experienced employees.

Sometimes much of it remains inside one person.

Some Decisions Need the CEO. Others Need What the CEO Knows.

One useful exercise is to track the questions and decisions that reach the CEO during a normal week and then ask why each one arrived there.

A portion genuinely requires CEO authority. There are strategic commitments, important relationships, and decisions that belong at that level.

But another group does not really require the CEO's authority.

It requires something the CEO knows.

Historical context. An understanding of priorities across departments. Knowledge about a customer relationship. The reason behind a policy. The ability to recognize that two situations that look identical on paper are actually different.

That distinction changes the diagnosis.

The organization may already have distributed work, responsibilities, and even significant decision authority. What it has not distributed as effectively is the context required to use that authority well.

Sales understands one part of a situation, Operations another, and Finance another. The company's systems may contain many of the relevant facts, but someone still has to understand how those facts relate to each other.

In many growing companies, that person is still the CEO.

Not because anybody deliberately designed the organization that way. It simply evolved that way because, for years, asking the CEO was easier than creating another mechanism.

Growth Can Make the Dependency Worse

This may explain why some CEOs hire stronger managers and add employees, yet their own workload does not decrease as expected.

Hiring creates more capacity to perform work, but growth also creates more interactions, exceptions, handoffs, competing priorities, and decisions. If many of the ambiguous decisions eventually depend on context concentrated in one person, adding capacity elsewhere does not necessarily remove the constraint.

The business can continue performing well for quite some time, which makes the dependency difficult to recognize. Customers are served, revenue grows, and problems are resolved.

The cost appears somewhere else.

The CEO starts earlier and finishes later. He checks something during dinner, answers a message on Saturday, or joins a call during vacation because this particular situation really does seem to require him.

No single interruption looks unreasonable.

Collectively, however, the company begins consuming hours that are invisible in any operating report.

I have come to think of those hours differently. In some cases, they are not simply evidence that the CEO works hard. They are compensating for something the organization has not yet learned to do without that person.

Part of the company's operating capacity is effectively being financed with the CEO's personal time.

At that point, this is no longer only a productivity problem.

It is a scaling problem.

The Answer Is Not to Document Everything

Trying to capture everything the CEO knows is unlikely to solve this. A mature company contains thousands of decisions, exceptions, relationships, precedents, and unwritten assumptions. Turning all of that into procedures would probably create more bureaucracy than clarity.

Judgment cannot be reduced entirely to rules.

A more useful starting point is narrower:

Which recurring decisions keep finding their way back to the same person, and why?

Sometimes the answer is authority. Sometimes information is missing. Sometimes departments are optimizing for different outcomes. Sometimes nobody remembers the precedent or understands the reasoning behind a policy.

And sometimes asking the CEO is simply easier than reconstructing the context elsewhere.

Those are different organizational problems. Calling all of them “delegation” hides the distinction.

The objective is also not to make every manager think exactly like the CEO. A healthy organization should be capable of reaching decisions the CEO might not personally have made, provided the reasoning is sound and the right context was considered.

That is a much higher standard than simply giving people permission to decide.

What I Ask Now

Today, when I hear a CEO of a growing company say, “I need to delegate more,” I take the statement less literally than I used to.

Sometimes delegation really is the answer. Sometimes the company needs stronger managers, clearer responsibilities, or a CEO willing to stop interfering after authority has been transferred.

But sometimes something more structural has happened.

The company has grown around a decision system that nobody intentionally designed because, in the early years, it didn't need to be designed. The founder remembered the customers, understood the history, connected information across functions, resolved exceptions, and knew why two situations that looked identical were actually different.

That way of operating may have helped the company succeed.

Eventually, the same dependency can make growth increasingly difficult.

I don't see this primarily as a technology problem, and I don't see it primarily as an automation problem. Technology may eventually become part of the answer, but starting there risks solving the wrong problem.

I would start with a question:

When you remove yourself from a decision, what exactly does the company lose?

If the answer is authority, perhaps the solution really is delegation.

But if the company loses context, history, judgment, priorities, or the ability to connect information scattered across different parts of the organization, then you may be looking at something deeper.

You may have built a company with capable people, defined processes, sophisticated software, and more data than ever before, while one of its most important systems still exists largely inside the people who built it.

Sometimes, inside one person.

And if that person is the CEO, working longer hours may keep the company moving.

But it will not make the company less dependent on them.