Every rank can escalate uncertainty, except one
At every rank in an organization, an uncertain decision has somewhere to go. A team lead who is not sure asks the department head; the department head who is not sure asks the VP; the structure works this way, and it is not a personal weakness of the person at any one rank — there is always a level above to carry the uncertainty, all the way up.
At the top, that structure simply ends. The CEO has no manager to ask. A board exists, but a board does not carry the daily reality of the business and usually meets monthly or quarterly, built for governance rather than for a weekly uncertain call. This is not a personality trait or a communication gap. It is the org chart running out, by design, at the very top.
This break is easy to miss because the person at the top is usually surrounded by people — an executive team, advisors, investors, an assistant. The crowd creates the illusion that an office above still exists. But none of those people is, by definition, above the CEO; they report or advise, and none of them formally takes on the decision.
This is a missing office, not a missing friend
Looking for an equal is aiming at the wrong target
Loneliness at the top usually gets named as "no one at my level to think with," and that is true but incomplete as a diagnosis. What is missing is not only an equal thinking partner; it is an office that formally takes the decision on, shares accountability for it — and that office cannot be filled by a good friend or a sharp advisor, because an office is a function, not a relationship.
The distinction matters because the same fix gets applied to both and neither works. A CEO looking to fix loneliness reaches for more friends, more network, more peer groups — all valuable, none of them doing what an office above does: formally sharing the decision, carrying part of the outcome, and having actual standing to question the call before it is made.
A peer group can tell a CEO "here is what I would do," but it does not carry the outcome of the decision. That difference looks small and is not — an office above draws its authority from actually being entitled to question the decision, and that entitlement is only built through a formal relationship, not an accidental friendship.
One example makes this concrete: a CEO can discuss a major layoff with a peer group; the group shows empathy, shares similar experience, and this genuinely helps. But once the meeting ends, the full weight of the decision sits back on the CEO alone, because the peer group is not and cannot become part of that decision. The same decision shared with a real office above works differently: once the office signs off, part of the outcome formally transfers to it too.
An office above draws its authority from being entitled to question the decision, not from offering advice.
Naming what used to go upward
The first step in filling this gap is not a general search for "someone to talk to." It is naming, retrospectively, which kinds of decisions used to travel upward. At earlier rungs of a CEO's own career, which decisions did they routinely send up — pricing risk, personnel calls, reputational exposure, major investment commitments?
That naming can be done with three questions:
- In my previous role, which kind of uncertain decision did I routinely send to the person above me?
- What did that person actually do beyond approving it — did they share the outcome, or only offer an opinion?
- Today, at the top, when that same kind of decision arrives, who do I send it to — or do I send it to no one at all?
Most CEOs stop at the third question, because the honest answer is usually no one. That answer shows exactly how concrete the gap is: not a vague feeling of isolation, but a specific, nameable function that has gone missing for a specific class of decision.
Answering these three questions is usually more uncomfortable than expected, because most CEOs have forgotten how often they once looked upward — it was natural and invisible then, the same way its absence is invisible now. Naming it is the first and only way to break that invisibility.
A designed office works differently from an accidental confidant
Once that naming is done, the work is not finding a general confidant but building a structure designed for that specific class of decision — meeting at a set interval, authorized specifically for that class of call, and genuinely sharing some part of the outcome. That structure is sometimes an expanded board committee, sometimes a standing outside-advisor relationship, sometimes a coaching cadence built specifically for this purpose.
The difference lies in frequency and authority. An occasional coffee is not there at the moment the decision actually arrives; a designed structure already is, because it was built for exactly that moment. What a CEO needs is not more people. It is a structure named, authorized, and scheduled in advance, matched to the type of decision it exists to carry.
The regularity of that structure matters, but so does whether the decision is genuinely shared. An advisor who checks in monthly but bears no accountability for the outcome has not become an office above — they remain a paid observer, and an observer, by definition, does not carry the decision.
This is a different, complementary layer to the equal thinking partner we describe in the piece on a leader's loneliness: one tests the thinking, the other formally shares the decision. The same person can sometimes do both, but they are not the same function, and one does not substitute for the other.
From the field
This pattern is not unique to one founder. Nearly every leader who reaches the top eventually runs into a class of decision they no longer send anywhere — the difference is when they notice it, and when they build something for it.
We once asked a founder a single question in a first session: in your previous role, who did you send a major uncertain decision to? The answer came instantly — the regional director. The second question took longer: today, as CEO, when the same kind of uncertainty arrives, who do you send it to? After a long pause, the answer was no one, and it seemed to be the first time they had heard themselves say it.
What we built with them was not a friendship. Together we named exactly which classes of decision needed to travel upward — major account losses, senior hiring risk, reputationally exposed statements — and set up a monthly, standing, pre-scheduled review for those three classes specifically. That structure was not friendship. It was a function, and building it meant accepting that the top itself creates a design problem, not that the founder was somehow weak — an acceptance that does its real work only once it becomes a routine rather than a one-time admission.
Six months later, that founder no longer spent time before a major decision wondering who to share it with, because the answer was already built in. Fixing loneliness at the top does not start with finding a person. It starts with naming which decision no longer travels upward. Requesting a conversation starts with doing that naming together.