The Accidental COO

This is a substantially revised version of a post originally written on October 28, 2024.

The first game ended quickly. I had been playing Chinese chess since boyhood, schooled by a gifted older brother. My new college roommate was good-natured about it. Then I won the second game, and his face flushed in a particular way — not quite anger, more the kind of humiliation that has less to do with being beaten than with seeing yourself beaten.

We had to share an apartment for the rest of the academic year.

I let him win the next two. Not conspicuously — just enough to give him something to stand on. I have never been able to fully untangle, in retrospect, whether that was a kindness or a quieter form of control. To choose deliberately to lose is to exercise the same power that crushing him would have exercised, only with a softer hand. For years I have remembered the move as an act of quiet generosity. I am no longer certain that is what it was.

What I am certain of is that the instinct it revealed — the instinct to make room for someone else rather than fill the room with yourself — would shape the rest of my working life. Power used generously multiplies; power hoarded quietly diminishes. The version of leadership that pulled at me, even before I had a vocabulary for it, was the one that treated stepping back as a form of strength. It would, eventually, define how I held the most consequential role I have ever stumbled into.

I never wanted to be a Chief Operating Officer. The title found me. For most of my career I was a developer, and I moved into management only when something around me seemed to be breaking and worth the trouble of fixing — though I will readily admit that "better" is always a matter of perspective. Once whatever I had set out to fix was fixed, I tried to step back. The limelight has never appealed to me.

Then someone close to me asked me to help him build a company from scratch. We called it illuminarean. His proposition was simple: he had ideas, and he wanted a team of talented software developers who could turn them into real products. The conviction I brought to the work was quieter — that the only way to attract and keep such a team was to build a culture in which the people came first, and to let the rest follow from that.

On paper the arrangement was unusual: in Korean corporate filings I was the person formally in charge. I had no interest in carrying the CEO title, and I operated, by every measure that mattered day to day, as the COO — a role for which I had no real template. He held the vision. I tried to hold the conditions that would let it survive.

What I learned, and what later research helped me articulate, was that the COO role does not lend itself to a single definition. Nathan Bennett and Stephen Miles, writing in the Harvard Business Review after dozens of interviews with executives, found that the role takes seven distinct forms depending on what the CEO needs at any given moment.¹ The most consistent thread across all of them was not a function but a relationship: a depth of trust between the two people at the top. Trust is the only thing that holds the role together. Everything else — strategy, operations, hiring — flows downstream of it.

The trust had to be earned, daily, and it had to be modeled. People watch what their leaders do more closely than they listen to what their leaders say. If we wanted the people we hired to give their best work to something they believed in, the work had to start with us.

What I came to understand, slowly and through accumulated mistakes, is that the most useful question a COO can ask is not what people can do for the company but what the company needs to do for them. The reframing seems small. In practice it changes the texture of nearly every decision — how you hire, how you run a one-on-one, how you decide who to promote and when, in difficult conversations, to let someone leave with their dignity intact.

The most important quality a leader can carry into a room, I have come to believe, is the willingness to ask, out loud, "I don't know. What do you think?"

It is harder than it sounds. There is a version of leadership — common, and often celebrated — that treats not knowing as weakness, and weakness as fatal. I have watched that version fail more times than I care to count, usually expensively. The more sustainable version, rooted in a genuine regard for others over oneself, turns out to produce better outcomes, more durable cultures, and, in the end, more meaningful work.

It also requires constant self-examination. A COO must take ongoing stock of their own capabilities and blind spots — must know when they are the right person to lead a conversation and when they are the wrong person to even be in the room. That awareness is, paradoxically, what makes psychological safety possible for everyone else. The leaders who build the strongest teams do not perform certainty. They create what the organizational behavior literature, following Amy Edmondson's foundational research, calls psychological safety: a shared belief among team members that the team is a safe place to take interpersonal risks — to disagree, to admit a mistake, to surface bad news before it becomes catastrophic.²

Edmondson studied fifty-one work teams in a manufacturing company and found that psychological safety was not merely a comfortable condition for morale. It was a primary enabler of learning. Teams that felt safe surfaced more problems, ran more honest post-mortems, and recovered faster from their own errors — a finding later replicated, at much larger scale, by Google's Project Aristotle, which studied more than a hundred and eighty of its own teams.³

Safety, however, cannot be installed by policy. It can only be modeled. It begins with the person at the top admitting that they do not have the answer, and meaning it. And it survives only as long as that admission is received, around the table, without consequence.

If you asked me what the work of a COO actually consisted of, day to day, I would tell you: hiring, listening, removing obstacles, and trying not to crowd the people growing around me.

The best COOs I have known, or studied, or observed from a respectful distance operate less like generals than like gardeners. A gardener does not force growth. A gardener prepares the soil, ensures the light, removes what is crowding the roots, and then, crucially, gets out of the way. In an organization that means investing in people before there is an urgent reason to, identifying the ones who might one day lead, and giving them room to stretch before any crisis requires it of them. In the early years of illuminarean, I came to think this kind of investment — unhurried, unpressured, genuinely interested in the individual — was the single most important thing I could do with my time.

McKinsey's Scott Keller and Mary Meaney, drawing on a decade of organizational research, argue that the companies most successful at retaining their best people are those that treat development as a discipline rather than as a reaction to attrition.⁴ Their findings track with my own, smaller experience. People who feel seen do not just perform better. They stay. And they recruit others like themselves.

The truth of this came home to me during the pandemic. The market for software developers was hotter than anything I had seen in years of building engineering teams. Companies were paying signing bonuses approaching annual salaries and still losing offers. We grew our team to thirty during that stretch. A CEO I had come to know through a university alumni group — whose small company had spent months trying to add a third developer to the two they already had — asked me, over coffee, how we were doing it. I did not have a tidy answer. What I had, it turned out, was a team that had been quietly doing the recruiting for me, vouching for the place by name to people they had worked with elsewhere.

An organization's culture is not what is written on the wall. It is the sum of what leadership does when no one is formally watching — or rather, when everyone is watching informally. Much of this work is unglamorous. A leader who treats no task as beneath them removes the hierarchy that makes others hesitate. Staying late on a tedious project, taking the first shift on something small, asking a junior employee how they are doing and genuinely waiting for the answer — these moments do more for an organization's culture than any all-hands speech. Commitment becomes something demonstrated, not demanded.

After everything — the strategy sessions, the board decks, the operational frameworks — I keep arriving at the same conclusion: hiring is the work. Most everything else is downstream of it. I made my share of mistakes in this area, and each one was instructive in its own uncomfortable way. Skills can be developed. Knowledge can be transferred. What is much harder to introduce after the offer letter is signed is values alignment — the degree to which someone shares the organization's sense of purpose, its instincts about how to treat people, its quiet standards of conduct when no one is watching.

An effective COO looks beyond the résumé. They look for people who will reinforce the culture when the leader is not in the room. They look for the kind of character that does not require an audience. I learned, the slow way, that hiring for credentials while overlooking character was a way of building a company that would not survive its own success — and that a single misaligned hire in the early stages can do more damage to a young organization's culture than almost any external setback.

I feel genuinely grateful to have worked with the people I did at illuminarean. They were principled individuals whose commitment brought something real into existence, and whose work, in many cases, is still bearing fruit somewhere I cannot see. They were the answer to most of the questions I did not yet know how to ask.

I want to be honest about something the leadership literature does not always acknowledge.

Servant leadership, gardening, psychological safety — these are not a hedge against every kind of failure. They are conditions that make good work more likely. They are not conditions that guarantee survival.

illuminarean closed in April of 2025. The proximate cause was external: financial pressure on our parent company that we could neither anticipate nor offset. The deeper question, which I have sat with for the better part of a year, is whether a different kind of leader — one with a more directive temperament and a sharper instinct for self-preservation — might have read the weather earlier and steered around it. I do not know the answer. I suspect it is partly yes, and I do not want to pretend otherwise. There were moments, in retrospect, when a harder edge would have served us better than the patience I tried to bring.

What I believe, on the other side of that experience, is something more measured than the version I might have written before the company closed. The kind of leadership I tried to practice produced real things: people who grew, work I am still proud of, relationships that have outlived the organization. It did not produce immortality, and I no longer think it was supposed to. A leader's responsibility is not to make their company last forever. It is to make sure that what gets built, for whatever time it lasts, is worth having built — and that the people who built it leave better equipped than they arrived.

I think often, now, of the people who worked with us. Their willingness to bring their full selves to something uncertain was the company's real product, and it did not vanish when the company did. They have gone on to other things. Some of them carry forward, in small ways I notice when we meet, the standards we tried to set together. That, more than any quarterly result, feels like the durable part of the work.

The chess game comes back to me now in a different light than it once did. I used to think of it as a parable about generosity — about choosing to lose so that someone else could find his footing. I am less sure of that interpretation. What I see more clearly is that the move was less about him than about the kind of person I was already trying to become, and the kind of room I wanted us both to live in.

That, in the end, is what the COO role asked of me, and what I think any second-in-command worth the title is asked. Not to win. Not even, in the most important moments, to be in charge. Only to make the room — and the work, and the people in it — worth the time we are given to share them.