The People Paradox

Two costs, one root

The Coordination Tax covered in the first chapter is the cost of how a firm spends its hours. Time as a resource.¹

The People Paradox, covered here, is the cost of how a firm fails to scale past its best people. People as a resource.

The two share a root cause, which is the absence of operational infrastructure. They land on different surfaces. The Coordination Tax shows up in the calendar: the IC memo that took three weeks, the quarterly report that took the better part of three weeks, the wire that did not clear because the bank account was not opened. It is a time problem. The People Paradox shows up in the org chart: the one analyst who knows where the data is, the one asset manager who can answer the LP question, the one CEO whose judgment has not been codified anywhere outside his head. It is a scaling problem.

A firm can pay the Coordination Tax and grow. Slowly, expensively, but it can grow. A firm cannot scale through the People Paradox. Past a certain size the best people become the ceiling. The firm tops out at the productive capacity of three or four critical individuals, and the next hire produces a fraction of the leverage the headcount math implies.

Marcus's quarterly report problem in Chapter 1 is the Coordination Tax in concrete form. Marcus's Priya problem, which this chapter walks into, is the People Paradox.

---

The archeologist, not the analyst

The Thursday before the data room request landed, Marcus stayed late.

Priya Atmani, the firm's investment associate, was the first person Marcus had hired who made him feel like the firm had crossed a threshold. She came from a $14 billion New York real estate private equity firm where she had spent three years as an acquisitions analyst out of undergrad. There, automated portfolio reporting assembled itself overnight, every IC memo had a template that loaded the deal's data already mapped, and decision logs wrote themselves because the system was designed to leave fingerprints. Priya was twenty-seven when she joined Chen Capital, fresh out of her full-time MBA at UT Austin. She had worked long enough at the New York shop to be fluent in how an institutional platform operates, but short enough that she had only ever worked inside one kind of system, and so could not name what she had been given. When she walked into Marcus's firm, she fitted herself to how things worked. Marcus had recruited her for eighteen months. When she accepted, he told his partners that Priya would change the firm's trajectory.

Fourteen months later, it was 8:15 on a Thursday evening. Marcus was walking past the open office on his way out. Priya was still at her desk. Both monitors lit. Three browser tabs, two Excel files, a Word document. The configuration of someone assembling rather than analyzing.

He stopped at her doorway.

"Walk me through what you're working on."

She looked up. "Quarterly. Fund II. The waterfall slide is the worst of it. The old version is somewhere in the shared drive, but the cap structure changed last quarter and I cannot find which version reflects that."

"How long have you been on it tonight?"

"Since six."

Marcus paused. "How long was it at your old firm?"

She thought for a beat. "We did not really do quarterlies the way we do them here. The package assembled itself. Every Sunday night a script ran, the data flowed in, by Monday morning the IR team had a draft they marked up and signed off on. Two days for ten funds and separate accounts." She caught her own phrasing. "I am sorry. That sounds like I am complaining."

"No," Marcus said. "Tell me more."

"It was like that for everything. IC memos had templates that loaded the deal data already mapped. Decision logs wrote themselves. When I joined here, I assumed every smaller firm did it the way you do it." She gestured at her monitors. "It is a different kind of work."

Marcus nodded once and let her go back to it. He took the elevator down without opening his laptop again.

---

Fifteen meetings, one source of truth

Tuesday, 8:30 PM. Marcus pulled up the calendar view on his laptop and ran the week side by side. Twelve meetings on his own. Eight of them with Priya. Her calendar carried nineteen in the same window. Two with LPs, one IC, one closing call. Fifteen internal meetings where she was the source of truth for something nobody had written down.

He scrolled to the shared drive and opened the Fund II LP reporting folder. Eight quarterly packages. Different file names. Different structures. Each one Priya had assembled alone.

He opened a draft message to her for Wednesday. Walk me through how much of your week is excavation. He stopped typing. He already knew. The firm had hired an analyst. The firm had been running on an archeologist.

He closed the laptop and left the draft unsent.

His phone vibrated on the desk.

you OK? Family Office just checked in on next steps and I haven't heard a word from you.

Sarah Kessler (capital markets advisor), from a car somewhere. He typed back.

Working on it. See you at 7.

That was the end of the exchange. The relationship was the kind where a handful of words covered what other relationships needed thirty minutes for.

---

Genius priced as assembly

Priya was still pulling occupancy data from the property management platform for the quarterly report. The system produced a PDF organized by property and month. What the report needed was a portfolio-level summary organized by investment thesis, vintage year, trailing-twelve-month NOI against original underwriting.

That view did not exist in any system. It existed in Priya's head.

The quarterly ritual: export from one system, reformat in Excel, cross-reference against the underwriting model, write narrative in Word, assemble into presentation.

Four quarters. Forty hours each. Two hours of analytical judgment. Thirty-eight hours of mechanical assembly.

Priya was the system because the firm had never built one. Her institutional knowledge made the assembly possible. It also made her irreplaceable. And it made her the bottleneck.

This is the People Paradox. The more capable the person, the more the system depends on them for work that does not require their capability. The more the system depends on them, the less time they have for work that does. The less time they spend on high-judgment work, the less the firm benefits from the capability it is paying for.

Because the capable person is always visibly productive, always essential to the next deadline, the CEO reads the constraint as a staffing gap. The answer looks like hiring. The answer is architecture. This is not particular to capital-intensive firms. Practitioners who systemize small teams report the same finding from the other end of the scale: the failure an owner reads as a staffing problem is, far more often, a defect in the process the staff were handed.² You are paying a premium for analytical talent and receiving administrative assembly.

---

I have watched or led the asset manager role at close range across four firms. The asset manager is hired to balance maximizing operating income, property value, and distributions. In practice, most of the week goes to what I call the goo in the system: cascading administrative work created by upstream handoffs that did not happen cleanly.

An LP asks when their K-1 will arrive. The asset manager cannot answer because accounting is waiting on a property report, because the property report is waiting on a sponsor response. Three emails to chase three different people, no time left for the trailing-twelve-month NOI on any single property.

An IR associate at a $1.8 billion fund had internalized the pattern completely. Her role was LP relationship management. What she actually did was build quarterly briefs by hand. Six hours per brief. Ten limited partners. Four quarters a year. Two hundred and forty hours of annual capacity routed into manual reconciliation: pulling distributions, cross-referencing positions, formatting charts, assembling narrative, writing the transmittal email.

The work was necessary. It was also the ceiling of her role. When the firm later built a platform that automated brief production, the per-brief time dropped to ninety minutes. Sixty hours per year. One hundred and eighty hours returned to LP cultivation: partner dinners, strategic calls about follow-on fund interest, deepening existing relationships.

The reclaimed capacity was the first moment the firm saw what she was actually capable of doing.

---

The career cost

The cost reaches beyond wasted hours into wasted career development. Every hour Priya spent reformatting quarterly data was an hour she was not building the pattern recognition that turns a good analyst into a great one.

The best people leave because they are under-utilized. The ones who stay stop bringing their full judgment to the room.

Hold periods have extended; every deal now generates more operational quarters than it did five years ago. The Coordination Tax per deal grows linearly with hold period. The team holds flat.

---

Why hiring does not solve it

An eighth person on a seven-person team running manual processes does not produce an eighth more output. In my experience, the eighth hire adds the headcount, multiplies the coordination complexity, and contributes only a sliver of net new productive work. Throwing bodies at a broken architecture multiplies the architecture's chaos.

Every new hire forces an apprenticeship the firm has never built. At firms without codified intelligence, the new hire walks into an absence and faces a political archaeology on top of the learning curve.

One CEO I advised brought in a capable associate to run underwriting. What he discovered in his first weeks was that underwriting lived as a series of judgments the CEO made in context, communicated verbally, adjusted in hallway conversations. The criteria existed nowhere on paper. They existed in the way different partners reacted to different deals.

He went looking for a decision log: a single artifact answering who decided what, based on what evidence, when, and with what dissent. Every firm should have one. This firm did not. He began building it himself, partly so he could do his job, partly because the artifact had to exist before his own work could become legible to anyone who came after him. Six weeks elapsed between his start date and his first substantive output. Three weeks mapping the decision topology. Three weeks convincing the partners that writing it down would make decisions repeatable rather than slower.

You cannot scale within a system whose logic you have not yet codified. Throw unharnessed AI on top of this and you accelerate the chaos.

---

The substrate, not the symptoms

Every visible problem this chapter has named (the asset manager's swamped Thursday, the IR associate's manual briefs, Priya's quarterly assembly, the senior hire spending his first weeks mapping hallways) is a symptom appearing on a surface the firm can see.

The source is somewhere else, in the operating substrate the firm has not yet built. Firms that work on the symptoms produce the next set of symptoms. Firms that work on the substrate find that whole categories of symptoms stop appearing.

---

Your COO already sees the structural gap

Claudia (CFO/COO) had been tracking this pattern for months before Marcus saw it.

She kept a running log of every operational item that escalated to senior leadership each week. On a single Monday in October, thirty-five items had crossed her desk. She sorted them into columns by who they belonged to.

Fifteen of them had no owner.

Not disputed ownership. No ownership. Items that landed on Priya because Priya was closest to the data, or on Jordan because Jordan was closest to the LP, or on Nathan because Nathan had touched the deal six months ago. None of those items was in any of their job descriptions. None appeared on any scorecard.

"These fifteen items belong to a function we haven't defined," Claudia told Marcus that Wednesday. "That's why everyone's stretched. They're not bad at their jobs. They're doing a job that doesn't exist yet on top of the job they were hired for."

Marcus had nodded, thanked her, and gone back to the pipeline. Claudia had filed the log and kept counting.

---

The second face

The fifteen unowned items carried a second diagnosis underneath the missing seats. Forty years of social psychology calls it diffusion of responsibility: the more people who could plausibly act on something, the less likely any one person does (Darley & Latané, 1968). Scale creates the disease. Every hire adds another they the team can route problems to. The analyst who spots a flawed assumption in a model but does not flag it because underwriting will catch it. The IR associate who notices a misstated figure in the quarterly letter but does not raise it because the CFO signed off. The asset manager who walks past the trash on a property tour because that is maintenance's job.

The bottleneck Priya carries is the first face of the People Paradox: the system depends on one person too much. The ownership gap Claudia tracked is the second face: the system depends on everyone, which is the same thing as depending on no one. SOPs and checklists route the work; they do not close the gap between noticing a defect and acting on it.

Here are two operating systems that do:

Disney trains every Cast Member, including executives walking the park in suits, to pick up trash on sight. Not a rule on a clipboard. An identity worn with the costume. There is no they because everyone is the they.

Toyota gives every line worker an Andon cord that halts a multimillion-dollar production line the moment they spot a defect. The radical part is not the cord. It is that pulling it is an obligation, not a right, and the team leader who arrives thanks the person who pulled it.

Neither pattern lives in the SOP binder. Both live in the operating system underneath. The Platform CEO's job is to engineer this at the firm's scale, to give a forty-person shop the felt ownership of a three-person founding team.

AI has added a third face to the paradox. When senior people move their drafting into private chat windows, the judgment juniors once absorbed by watching goes dark. The analyst no longer sees the managing director mark up a memo; the markup happens in a window of one. A firm can train its platform and starve its apprentices at the same time, and most firms doing the first are doing the second without noticing. The fix appears in Chapter 11: run the work in the open, where the corrections that teach the platform also teach the people.

---

The shape before the words

He thought about Priya's Thursday night. Forty hours per quarterly report. Thirty-eight of those hours were mechanical. Her analytical expertise got two hours per quarter. A new analyst would not change that ratio. Within six months, they would be spending forty percent of their week on coordination overhead. He pulled up the job listing that landed Priya fourteen months ago. It codified exactly the kind of person who would walk in on day one, ask where the reporting templates were, learn within a week that the answer was "there are not any," and then figure it out. What happened? Or more importantly, what did not happen?

Priya was a bottleneck because the firm had made her the system. Remove the coordination overhead and Priya was a principal-track investor with institutional training, fourteen months of firm-specific knowledge, and judgment that would mature into something irreplaceable. The firm did not need another Priya. It needed a platform that let the Priya it already had do the work she had been hired to do.

He thought about Nathan Park, head-hunted from a top competitor to run acquisitions because the firm wanted his deal instinct, then placed at a desk where most of his week went into a spreadsheet he had built himself because the CRM could not produce the view he needed.

He thought about Scott Engel, the firm's Head of Asset Management with fifteen years in the business, who had spent three weeks last quarter assembling the board's performance report rather than writing it, pulling occupancy data and reconciling against Tom Langford's (controller) accounting output. Both hired for judgment. Both buried in mechanical work the firm had never built the architecture to absorb. And underneath both, the quieter cost Claudia had been counting for months: the items nobody picked up because everyone could have. David Kwon, the senior asset manager who ran the harder half of the portfolio under Scott, was buried the same way after nearly a decade at the firm.

He did not have the words to describe it, but its shape was already clear.

---

A question Priya had been holding

Wednesday morning. Marcus stopped by the office early on his way to the Driskill. Priya was already at her desk, which surprised him. She was reading something on her left monitor and writing notes by hand on a legal pad, which surprised him more.

"You are in early," he said.

"There is something I have been meaning to ask you for fourteen months. Now seems like a moment to ask it."

Marcus set his bag down.

"The IC memo template," she said. "The one that says what sections every IC memo carries, what the downside scenario has to argue, who has to sign off on the rent-growth comp. Is there a reason it does not exist, or has it just never been built?"

Marcus looked at her for a long moment.

"It has never been built."

"OK." She made a note on the pad. "If I drafted one this week and showed it to you Friday, would you read it?"

"Yes."

"Good."

She turned back to her monitor. Marcus picked up his bag and walked out toward the elevator. He had twenty minutes until breakfast with Sarah. He spent the drive thinking about what he had just been handed.