The board meeting where everything changed did not feel like the board meeting where everything changed. It felt like every other board meeting. The room was the same. The coffee was the same temperature it always was, which was slightly too cold because someone always made it slightly too early. The slide deck had the same structure it always had, the quarter in review, the metrics, the outlook, the risks, the questions. The people around the table were the same people who had been around the table for the previous six meetings, sensible people who understood what was being built and who had, until recently, been patient about the pace of it. The patience had a shape. The CEO had known this from the beginning. Investor patience is not an absence of urgency. It is urgency deferred. The questions are still there. The expectations are still there. The timeline is still there, ticking quietly in the background of every conversation, present in the careful phrasing of every question that sounds like curiosity but is really a measurement. At the sixth board meeting the patience ran out. Not loudly. Not with an ultimatum or a confrontation or any of the things that make a moment easy to identify in retrospect as the moment everything changed. Just a question, asked by Marcus, who was the kind of investor who had made his money by being the first person in a room to say the thing everyone else was thinking but had not yet decided to say out loud. "At what point does the growth rate justify the cost structure?" asked Marcus. The question was precise. It had been prepared. The careful neutrality of the phrasing, at what point, rather than why has not, was the kind of precision that comes from someone who knows exactly what they are asking and has chosen the least confrontational version of it. The CEO answered. The answer was honest and accurate and did not satisfy Marcus, not because it was a bad answer but because the question had not really been a question. It had been a signal. The signal was: the current approach is sustainable up to a point and we are approaching that point and the board would like to understand what happens next. What happens next. The CEO drove home that evening thinking about those three words.

There was a light on in the kitchen when the car turned into the street. There was almost always a light on in the kitchen. The person who lived in the kitchen light had been there for eleven years and had learned, over those eleven years, the specific rhythm of the CEO's late returns, and had learned also that the late returns were not all the same, that some of them were the lateness of productive exhaustion and some of them were the lateness of something heavier, and that the kitchen light was the right response to both even though it could not distinguish between them. The CEO sat in the car for a moment before going inside. The street was quiet. Two houses down a dog was barking at something and then stopped. Inside, the kitchen light was warm and yellow. There was food on the stove, the lid on the pan indicating it had been kept warm rather than freshly made, the consideration of someone who had prepared for the lateness without requiring acknowledgment of having prepared. They did not talk about the board meeting that evening. Not because the conversation was avoided. Because the kitchen, and the food, and the light, and the person who had kept it warm, were themselves a kind of conversation, and the CEO was grateful for it in the specific way that you are grateful for things that ask nothing of you in the moments when you have nothing left to give. The company had grown in a way that felt, from the inside, like controlled acceleration. Each new phase had been prepared for before it arrived. Joel's systems had scaled without breaking. Maya's knowledge base had grown without losing its usefulness. The team had expanded from twelve to forty-seven people and the forty-seven people felt, in most of the ways that mattered, like an extension of the original twelve rather than a replacement of them. This was not an accident. It was the result of a hiring process that valued the right things and a culture that transmitted itself through the way people worked rather than through the documents that described how people were supposed to work. The CEO had maintained the habit, mostly, of reading the support tickets. Of walking over to the desk of the person who had done something right. Of sitting with Maya or Joel without a specific agenda, just to stay connected to the texture of the work. Mostly. The gap between mostly and always had been growing for about eight months. The CEO knew this the way you know a thing that you have chosen, not consciously but functionally, not to look at directly. The support ticket reading had gone from daily to weekly to something that happened when there was time, which meant it happened less and less because there was always something that seemed more urgent and the tickets were quiet and the quiet things always lose to the loud things in the competition for a CEO's attention. The walk to Priya's desk had not happened in four months. This was not a catastrophe. The company was still working. The culture was still visible. The customers were still writing the kinds of things in their feedback that suggested the person who had helped them had been a person and not a process. But the CEO had stopped being the one who knew this from direct experience and had started being the one who knew this from reports, and there is a difference between those two kinds of knowing that does not show up in the reports themselves. Marcus's question had made the difference visible. At what point does the growth rate justify the cost structure. The cost structure had three main components. The product team, which was producing the things that made the company worth using. The operational team, which was making sure that using it actually worked. And the support function, which was the part of the company that customers experienced most directly and which had, until recently, been the source of the kind of feedback that made the CEO feel certain about the direction. The support function was expensive. Not wasteful. Expensive in the way that quality is expensive, which is to say that it cost more than the alternative and produced more than the alternative and the difference between the two was real but hard to put in a slide deck in a way that made the investment look obvious rather than sentimental. Marcus was not a sentimental investor. Callum had joined in the second year, in a role that required him to work closely with Yusuf, who had been with the company since the end of the first year and who had, in the way of people who arrive early enough to feel ownership of a thing, a strong sense of how the work should be done and a correspondingly strong reaction to people who did it differently. The friction between them had started quietly. The way friction always starts. Not with a confrontation. With a series of small moments in which two people discovered that they had fundamentally different approaches to a shared task and that neither approach was wrong in absolute terms but that both approaches were wrong when they were happening simultaneously without coordination. Sandra had managed it. Not by resolving it. By containing it. By making sure that Callum and Yusuf were not assigned to the same tasks without a structure that gave them separate lanes, and by having the individual conversations that her instinct told her were necessary before the friction became visible to the people who did not need to see it. The conversations were the kind that Sandra was good at. Specific. Without blame. Oriented toward what needed to change rather than toward what had gone wrong. Callum had responded to them. Yusuf had responded to them. Neither of them had changed in the fundamental way that would have resolved the friction rather than containing it, but the containment had worked well enough that the CEO had not been aware of the situation except as a vague background awareness that Sandra was managing something in the team in the way that Sandra was always managing something in the team.

Hana and Tomás had been working together for two years by the time the board meeting happened. They worked well together. They had the particular efficiency of people who understand each other's working style without having had to explain it, who know when the other person needs space and when they need input and who give each of those things without being asked. They were not friends in the way that required socialising outside the office. They were something more specific than friends and less defined than anything that had a clear name. On the Thursday evening after the board meeting, when most people had gone home and the office had the specific quality of a room that has been full of people and is now almost empty, Tomás was still at his desk. Hana was still at hers. They were not working on the same thing. They were not talking. But there was something in the shared quality of their staying, in the fact that they had both been there at seven and were both still there at eight, that felt like a kind of communication even in the absence of words. Hana looked up at some point and found Tomás looking at her. He looked away immediately. Back at his screen. She looked at the small dark-green cactus on the corner of her desk, the one that had survived four moves and one winter with no heating, and she thought about the first time, and the Spanish wine with the red and gold label, and the way he had looked around the room as if committing it to memory. Then she picked up her bag and said goodnight. "Goodnight," said Tomás. He did not look up from his screen. She walked out. He stayed. Nothing happened. But it was closer than before. The consultant arrived three weeks after the board meeting. Not invited by Marcus directly, but suggested in the way that things get suggested in board conversations, obliquely, by reference to a firm that had done good work for a company Marcus knew, mentioned in the context of a broader conversation about operational efficiency, allowed to land and not contradicted, which in board dynamics is close enough to an instruction. The consultant's name was Richard. He was competent, well-dressed, and had the particular confidence of someone who had been in enough rooms to know that confidence itself was a deliverable. He asked good questions. He listened to the answers with the focused attention of someone who is simultaneously taking notes and forming conclusions, the kind of listening that is productive but not quite present, the way a doctor listens to symptoms while already moving toward a diagnosis. The CEO showed Richard everything. The systems. The knowledge base. The support operation. The metrics. Richard was shown Priya, not by name, but as an example of the kind of person the support team was built around, the instinct to stay with a problem until it was sorted because leaving it unsorted was not something that person considered. Richard nodded at this the way he nodded at most things, with a considered expression that communicated neither agreement nor disagreement but simply the registration of information received. At the end of the second day Richard asked if the CEO had considered an outsourced model for the support function. The CEO said: considered and decided against. Richard said: on what basis. The CEO said: the support function is where our customers experience us. The people doing that work carry knowledge that took years to build. The culture that produces a Priya is not something you can specify in a service level agreement. Richard said: I understand that concern. I would push back on the assumption that it cannot be managed through a well-structured outsourcing arrangement. The providers have improved significantly. The contractual frameworks have become more sophisticated. The cost differential at your current scale is significant. The CEO said: how significant. Richard said a number. The number was larger than the CEO had expected. Not because the CEO had not known the support function was expensive. Because the number Richard said represented what could be saved annually, and when a number that size is said in a room the room changes slightly, the way a room changes when someone opens a window and the air shifts, and everything that was true before the number was said is still true but sits differently in relation to it. The CEO said: I will think about it. Richard said: of course.

The thinking took six weeks. The CEO talked to Maya, who said what the CEO already knew, that the knowledge in the support function was real and the culture that produced it was real and that both of those things would be at risk in an outsourced model regardless of how well the contract was written. The CEO talked to Joel, who said that the operational systems could be integrated with an external provider but that the integration would require work and that the quality of the output would depend on how seriously the provider took the integration and that in his experience providers took integrations seriously until they did not, which was usually around the eighteen month mark when the contract renewal conversation began to feel more important than the operational detail. The CEO talked to Sandra, who had built the support team from six people to thirty-one and who knew every one of those thirty-one people well enough to know which ones needed support on difficult calls and which ones were ready for more responsibility and which ones would be lost immediately if the stability of the team changed. "If you do this, you will lose Priya," said Sandra. The CEO said: Priya would be offered a position with the provider. "Priya would leave," said Sandra. "She is not the kind of person who does well inside a large outsourced operation. She does well here because of what here is. Change what here is and you change what Priya is able to be." The CEO thought about this for a long time. Then thought about the number Richard had said. Then thought about Marcus's question. At what point does the growth rate justify the cost structure. Then thought about the next board meeting, which was in four weeks, and about the slide that would need to show a credible path to the kind of margin that would make the growth rate and the cost structure sit comfortably next to each other in the same deck. The thinking that followed was not dishonest. The CEO did not tell a story that was not true or ignore information that was inconvenient. The thinking was the ordinary kind of thinking that happens when a real constraint meets a real cost and the person doing the thinking is under real pressure to find a solution that works on paper in a room full of people who are measuring the paper and not the thing the paper is describing. The CEO made a list of the risks. Sandra's concern about Priya was on the list. Maya's concern about the knowledge transfer was on the list. Joel's concern about the eighteen month mark was on the list. Against each risk the CEO wrote a mitigation. Transition plan. Knowledge transfer process. Integration protocol. Contract clauses around cultural alignment and knowledge retention. Service level agreements that specified not just the what but the how. The mitigations looked credible on paper. The CEO had been in enough rooms to know that things that look credible on paper have a specific relationship with reality, which is that they describe what should happen rather than what does happen, and the distance between those two things is where most of the risk actually lives. The CEO knew this. And then the CEO went to the board meeting and presented the outsourcing proposal and the number Richard had said and the mitigation plan and answered Marcus's questions and watched the board approve the decision with the satisfied expression of people who have been waiting for a particular answer and have finally received it. On the drive home the CEO thought about Sandra's words. She is not the kind of person who does well inside a large outsourced operation. She does well here because of what here is. The CEO thought about Priya at her desk, headset on, staying with a problem until it was sorted because leaving it unsorted was not something she considered. The CEO thought about Owen saying I am in the room that matters. The CEO thought about Nora saying ask me again in the spring. The CEO thought about Hana and Tomás and the Spanish wine and the cactus on the corner of the desk. The CEO thought about the green notebooks and the systems Joel had built before they were needed and the room above the dry cleaner where the whole thing had started with three people and one idea that would not leave them alone. Then the CEO went home and the light was on in the kitchen and the food was warm on the stove and nothing was said about any of it, not that evening, not in the way of avoidance but in the way of a relationship that has learned when the right time for a conversation is and that this was not it. The CEO ate. Sat for a while. Went to bed. The decision had been made. The living with it was about to begin.

End of Chapter 3

Writer's Thought:

The CEO heard everything Sandra said. Then went to the board meeting.

Here is What is Broken. The CEO. The Pressure.

MarvinPro | March 2026

marvinpro.com

Think Simple.