The Paleo Workplace — Introduction & Outline

Int.

Introduction

Two Rooms

“You pay for culture, or you pay for turnover. Every time.”

I was thirty-five, unemployed, and scared, and I was in the lobby of 555 California Street interviewing at a law firm.

I had just walked away from a good job. Regional network administrator for a national employment law firm, four cities, real responsibility. I’d quit because my bosses three thousand miles away had made a sport of ambushing me on conference calls, and one morning I told four of them at once that I was sharper than all of them combined and hung up. I meant it. I was also, within a few weeks, out of work with no plan, and I did what a frightened animal does. I ran back toward the grid.

So there I was in a marble lobby, waiting, looking down a corridor at a row of offices. One of them was empty — a young attorney, out at lunch. And I could see the whole of his working life through that doorway. A box, maybe ten feet by twelve, with a window. Books. A monitor. A stack of documents in the middle of the desk.

And on the shelf, a photograph of his wife and their child.

That photograph was the reason he was in the box. It was also the only thing in the room that had anything to do with being alive.

I knew exactly what his day was, because I had done a version of it. I’d gone back to school in my thirties for a paralegal degree and spent years inside the machinery of litigation. Parsing. Wordsmithing. Doing math with words. Reading four hundred pages to find the one sentence that mattered and then arguing about the comma in it. It is genuinely difficult work and there are people built for it, and I was not one of them. It made me drink. It made me lean hard on coffee and appetite and whatever else got me to Friday.

He couldn’t have liked it either. Nobody likes it. But standing there I understood, with no judgment at all, why he was doing it anyway, and the answer was sitting on the shelf in a frame. He had dependents. Somebody was counting on the health insurance and the mortgage payment and the predictability. He may as well have been wearing leg irons. An open-air prison with a partner track.

Here is the part I want to be honest about, because it’s the reason this book exists in the shape it does. That was how he looked to me when I was flush. Later, in the lean years — and there were lean years, because I’ve started a dozen businesses and most of them failed — I thought about that same man and saw something completely different. I saw somebody who had been smarter than me. Somebody with a salary and a title and a family who had never had to explain to anyone why the money was late.

Same man. Same box. Same photograph. The only variable that moved was my bank balance.

So this is not a book that tells you to quit your job. I don’t have standing to tell anybody that, and anybody who does is selling something. This is a book about the fact that a competent, educated, well-paid adult should not have to choose between loving his family and doing work that doesn’t hollow him out — and that the choice exists not because of him, and not because of his employer’s malice, but because of the box.

You already know what a mismatch looks like when it isn’t us. A giant panda in a good zoo has veterinary care, no predators, and more bamboo than it can eat, delivered forever. By every metric a spreadsheet holds, it’s better off than its cousins in the mountains, who are cold and hunted and occasionally starving. And it will not breed. The elephant sways in place for hours. The parrot pulls out its own feathers. Nobody blames the animal, suggests a resilience workshop, or asks it to be more grateful for the bamboo. Everybody understands that you have taken a creature out of the ecosystem it was built for and it has stopped working right, and the fix is the enclosure and not the animal.

Then Monday comes and we sit in a box under fluorescent light, doing a fragment of a job we will never see finished, waiting on approval from someone who has never done the work, and we call the resulting misery a personal failing. We put it on the individual. We send him to a webinar.

Now let me show you the other room.

Some years later — flush this time, running my own shop — I walked onto the campus of a technology company that had more money than it knew what to do with. A unicorn, mid-boom, hiring hard and losing sleep over how to keep brilliant people from wandering off to the unicorn across the street.

The exterior wall was glass, floor to ceiling, so the whole floor sat in daylight. There was a fountain inside the building — moving water you could hear from anywhere in the room. Plants. Long open sightlines across the entire group, so you could see who was there and what was happening. Food, free, laid out in the open, eaten together. Massages, if you wanted one. Twenty percent of your week to spend on whatever you thought was worth doing. Work from home when it suited you. One chief at the top and almost nothing between him and anybody else. You could walk up to any person in the building and talk to them. You had a task, and how and when you did it was your business.

If you took a person from fifty thousand years ago and asked them to design a place to work — the architecture and the management both — this is what they would draw.

Sunlight. The sound of water. Green things growing. A clear view across your whole band. Food shared in the open. One chief, and no layers between you and him. The freedom to talk to anyone. A task that is yours, and discretion over how you do it. Children and play mixed in with the work rather than quarantined from it.

That is not an office. That is a camp by a river.

And here is the part that made me start writing. Nobody there was trying to do that. Nobody at that company had read a book about hunter-gatherers or given a moment’s thought to the Pleistocene. They had unlimited money and a serious, expensive problem — keeping people who could leave — and they solved it by removing constraints one at a time. Cost. Convention. The received wisdom about what an office is and how a company is supposed to behave. And when enough constraints came off, this is what fell out. They didn’t design a Pleistocene camp. They simply stopped preventing one.

That, to me, is stronger evidence than anything in a journal, and it’s the whole reason for this book. It’s a revealed preference. Here is what the richest and most talent-starved organizations on earth build when expense stops being the deciding factor — not a theory about what humans need, but a demonstration, run at enormous cost by people with no thesis to prove.

NOTE — One feature on that campus did not arrive by removing constraints. The open floor plan descends from Taylorist bullpens and exists to fit more bodies into less square footage — and it is the single element on the list with genuinely bad evidence behind it. The mechanism predicts that. A camp is a small band with acoustic privacy, individual shelters, and the ability to walk off. It is not four hundred people in one uninterrupted room. See Ch. 21.

So we can stop guessing at the specification. Read it off the campus, then check each line against research on living human beings in real jobs, and it holds up item by item. A whole piece of work you can point at and call yours. Authority over your own method. Visible progress. Recognition inside a group small enough that your contribution can actually be seen. Variety. A real share of what you help bring down. None of it is exotic, and none of it requires you to believe anything in particular about the Pleistocene — the ancestral story explains why the list looks like this, not that it does.

Now put that list against what a large organization actually issues you. A fragment of a job, with the outcome three departments away. A method dictated by someone with less expertise than you. Progress measured quarterly, if at all. Recognition distributed by a manager who supervises forty people and knows your name and nothing else. Salary that pays the same whether the thing worked or not. That isn’t a bad company. That’s just a company. It is the standard-issue enclosure, and it violates nearly every line of the specification.

Two rooms, then, built in the same century by the same species, both called a workplace. The box at 555 California is what we build when cost and convention are in charge. The campus is what we build when they aren’t. Almost everybody works in the first one.

The mismatch isn’t a mood. It shows up in mortality data, in turnover, in healthcare spend, and in the innovation that never happened because the person who’d have had the idea was busy getting permission.

Which raises the obvious question: if it’s this bad, why has nothing changed?

Because it isn’t this bad. Not visibly. The arrangement wasn’t imposed on anyone — it evolved, over about a century, under conditions where there genuinely weren’t many alternatives, and it fit those conditions well enough. Payroll runs. Product ships. Buildings get built and drugs get approved and nobody riots. By the standard of a system that has to not collapse, the modern workplace is a success.

It sort of works. That is exactly the problem. A thing that fails outright gets fixed, because failure is legible and somebody has to answer for it. A thing that sort of works gets defended for a century, because the gap between what it delivers and what it could deliver never appears on any statement anyone reads. Nobody line-items the campaign that was good instead of great, the engineer who left in year three, the idea that died in a review cycle. The waste is real, it is enormous, and it is structurally invisible. That’s not a scandal. It’s an accounting problem, and accounting problems are the most durable kind.

One thing has shifted underneath it, and it’s worth naming before we start. The cost of building something — a product, a service, an audience, a company — has fallen off a cliff in about fifteen years. Taylor Pearson made that case to the individual worker in The End of Jobs, and made it well; I’m not going to re-argue it. What follows from it for an employer is simpler and less discussed: you are no longer competing only against other employers for talent. You’re competing against the option of not having an employer at all. Most people won’t take that option. But the ones who could are, almost by definition, the ones you most want to keep.

I should say plainly who’s telling you this, because the credential is unusual and it isn’t a degree.

I’m fifty-two, and I’ve been working for thirty-six years. I finished my coursework at Berkeley in 1996 and did not receive the diploma until 2005, when I finally paid Campus Accounts Receivable the three thousand dollars I owed them. For nine years I had the whole education and none of the paper. What I had instead was a line on a résumé that read University of California, Berkeley, and it turns out the line is what people hire. Nobody ever asked to see the document. In 2003 I needed a credential to get into an ABA-approved paralegal program and didn’t have one, so I went back to San Diego Miramar College and had them retroactively award me an associate’s degree for coursework I’d finished years earlier. I used the AA to get in. The BA was still sitting in an office in Berkeley with a hold on it.

Being biracial helped, and I’ll say that plainly rather than dance around it. A clean-cut, well-spoken Black kid with Berkeley on his résumé got a particular kind of welcome — people wanted to be part of the story going well, wanted the hire to work out. I’ve never resented that and I’m not going to apologize for it, because it was also justified. I was highly competent, every single time, at whatever they put in front of me. The door opened a little easier. What happened after it opened was mine.

And it opened a lot. I have held something close to a hundred jobs in thirty-six years. Grossing technician in a pathology lab, which is exactly as pleasant as it sounds. Receiving at ZymoGenetics, in the old Seattle Light and Power building on Lake Washington. Summarizing case law for JAMS. Colorado National Bank. Nike. Amazon. E-discovery databases, then networks for four cities. And in the lean stretches, because there were plenty of lean stretches: building sandwiches at Subway, digging ditches, installing cultured marble. There are jobs I took for a few months between consulting contracts that have never appeared on any résumé and never will.

I was Jared from The Pretender — dropping into a new profession, working out the profession, moving on. Show me the job and I’ll learn the job.

And almost every time, the same thing happened. It started to suck, and I left.

I want to be precise about “it started to suck,” because it was never the labor. I came into every one of those jobs open-hearted. I liked the people. I was good at the work and I wanted to be good at it. And then, somewhere between month four and month twenty, the rug came out. An ambush call. A decision that belonged to me handed to a room. A manager who needed me smaller than I was in order to feel correctly sized himself. Recognition that went sideways to someone who hadn’t done it. Nothing dramatic and nothing actionable — just the moment you see the shape of the thing you’re actually inside, which is not the thing you were told you were joining.

The only reason my story diverges from that attorney’s is that I never signed in blood. No mortgage I couldn’t walk away from. No kid depending on the insurance. When the rug came out, I could stand up, and I did, over and over, for three decades — not out of courage, but because the bill for leaving was one I could personally cover.

Add it up and you get something no ordinary career produces. Pathology lab, shipping dock, bank, national law firm, global sportswear brand, one of the largest retailers on earth, a two-man startup, my own consultancy, and a long tail of jobs nobody would put on a wall. Blue collar and white. Hourly, salaried, contract, self-employed. Fortune 50 and four people in a room. Most people get two or three work environments in a lifetime and quite reasonably draw their conclusions about work from a sample of two or three. I have a sample in the dozens.

I didn’t set out to be a student of work environments. Quitting made me one.

Because after the twentieth or thirtieth time you watch the same failure play out — different industry, different people, different product, different pay grade, identical shape — you stop believing it’s about the industry. The pathology lab and the law firm and the sportswear giant were running the same operating system and producing the same defects, and none of them could see it, because each one thought its problems were its own. I got to see the pattern underneath because I kept moving. I saw the Matrix in code. This book is me telling you what’s in it.

What follows runs in five parts. Part I is the specification and the gap — what the animal needs, what the box supplies, and what the difference costs in bodies and in money. Part II is a short history of how the box got built, because it was assembled for reasons that made sense at the time and it helps to know them. Part III is what has changed underneath it. Part IV is the build: what you actually do, at scales between four people and forty thousand, on a budget that isn’t a unicorn’s. Part V is whether it pays, including the places this has been tried and failed, and the numbers from my own companies — the ones that didn’t work as well as the ones that did.

Steve Jobs used to say that if you look in the mirror too many mornings in a row and don’t want to do what you’re about to do, something needs to change. Most people hear that as career advice, and as career advice it’s expensive and not available to everybody. I’d like to offer it as something else: a research method. I took it — not because I was brave, but because I could afford to — and I took it for thirty-six years across a hundred jobs.

This book is what you get to see if you keep quitting.

part i · the animal and the box

The Man in the Box

The 555 California lobby, in full. The empty office, the photograph, the recognition of why he stays — and the honest admission that the same man looked like a prisoner when you were flush and like a genius when you were broke. That flip is the book’s moral license: it earns the right to criticize without smugness.

SOURCE — Original. Establishes the enclosure, the lock, and the author’s ambivalence in a single scene. Do not resolve the ambivalence here or anywhere — it’s load-bearing through to Ch. 28.

Fifty Thousand Years Old

The campus, in full, as the book’s central piece of evidence. Sunlight, moving water, greenery, sightlines across the whole band, shared food in the open, one chief with nothing stacked beneath him, freedom to approach anyone, a task and discretion over how you do it. Nobody was trying to build a Pleistocene camp. They had money, a talent problem, and permission to remove constraints — and this is what fell out. A revealed preference, not a theory.

SOURCE — Original, and the origin of the whole project. SUPPORT — Laszlo Bock, Work Rules!, for what Google’s People Ops was actually optimizing and with what data · Google’s Project Aristotle · the 20% time literature and its later retrenchment, reported honestly · the design-brief record on tech campuses of that era. Note the one feature that didn’t come from removing constraints — the open floor, inherited from Taylorist bullpens — and that it’s the one with bad evidence behind it. The mechanism predicts that; it’s a confirmation, not a hole.

What the Animal Requires

The specification, derived from Ch. 2 rather than asserted: a whole piece of work with your name on it; authority over your own method; visible progress; recognition inside a group small enough to see you; variety; a real share of the kill. Then check each line against the research. Engineering tolerances, not values.

EVIDENCE — Hackman & Oldham, Job Characteristics Model (1976) — skill variety, task identity, task significance, autonomy, feedback · Deci & Ryan, self-determination theory, and the 1999 meta-analysis on control displacing intrinsic motivation · Teresa Amabile & Steven Kramer, The Progress Principle (2011) · Gallup’s Q12 items, read as a description of a small band · the professions that top satisfaction surveys (clergy, firefighters, physical therapists) and what they share · the biophilic design literature for the physical half: Ulrich, Kaplan & Kaplan, Appleton, Terrapin’s 14 Patterns.

What the Box Supplies Instead

Run Ch. 3 in reverse, line by line. A fragment instead of a whole job. A dictated method. Quarterly feedback. A manager of forty. Pay indifferent to outcome. Not a bad company — a standard one, which is the point, and the reason nobody fixes it.

EVIDENCE — Gallup, State of the Global Workplace (engagement flat and low for a decade; manager engagement itself now declining) · Microsoft Work Trend Index on meeting load, after-hours creep, and productivity paranoia · Cal Newport, A World Without Email and Slow Productivity, on pseudo-productivity · Bernstein & Turban (2018) on open-plan conversions cutting face-to-face interaction.

The Lock

Why intelligent people stay in arrangements that are hurting them. Not docility — arithmetic. Dependents, mortgage, employer-tied health coverage, vesting schedules, and a cost of living that rises to meet the salary. The lock is why the mismatch persists without anyone having to enforce it.

EVIDENCE — Job lock research on employer-sponsored health insurance and reduced mobility (Madrian and the successor literature) · household debt-service ratios · golden handcuffs and vesting-cliff design · Mullainathan & Shafir, Scarcity (2013) — money scarcity consuming cognitive bandwidth, which is the bill on the other side of the trade · original: the photograph on the shelf as the cleanest statement of the mechanism.

The Body Keeps Score

The mismatch is not a feeling. It is measurable in physiology, and the specific toxin is not hard work — it’s demand without control. This is where the panda stops being a metaphor.

EVIDENCE — Michael Marmot, The Status Syndrome and the Whitehall studies — mortality gradient by rank, mediated by job control; the hardest single fact available to this book · Robert Sapolsky, Why Zebras Don’t Get Ulcers · Karasek & Theorell demand-control model · Jeffrey Pfeffer, Dying for a Paycheck (2018) · WHO’s 2019 classification of burnout as an occupational phenomenon.

The Ledger

Itemize the corporate bill, and explain why none of it appears on a statement anyone reads. This chapter is where a CFO either joins you or closes the book, so it should be the most numerically disciplined thing in the manuscript — real figures, not the round ones everybody repeats.

EVIDENCE — Gallup Q12 meta-analyses linking engagement to profitability, quality defects, safety, and turnover at business-unit level · Gary Hamel & Michele Zanini, Humanocracy (2020) — bureaucracy cost estimate and the bureaucratic mass index · SHRM turnover replacement costs · healthcare spend attributable to workplace practice (Pfeffer) · the accounting argument itself: costs that land in a different budget line than the decision that caused them.

The Whole Job, Broken Up

One person used to make the whole thing. Industrialization split that person into fragments because fragments were more efficient — and created, as a byproduct, a job whose entire content was coordinating the fragments. Custody didn’t get stolen. It got engineered out, for good reasons, by people optimizing something else.

EVIDENCE — Adam Smith’s pin factory, read for the cost rather than the gain · Taylorism and the deliberate separation of planning from doing · Hackman & Oldham’s task identity as the precise thing specialization removes · the honest concession that some production has to be mechanized, and a way for the reader to locate their own work on that spectrum · cross-reference: Paleo Project Management, Ch. 16.

The Manager Is a Fossil

Middle management answered a real coordination and communication problem. That problem has been substantially solved by other means, the role persists on inertia, and the market has recently begun removing it without anyone framing it as a philosophical shift.

EVIDENCE — Ronald Coase, “The Nature of the Firm” (1937) · Alfred Chandler, Scale and Scope — the honest opposing case, conceded where it holds · Bayer’s Dynamic Shared Ownership restructuring (2024), including the rulebook reduction · Andy Jassy’s September 2024 Amazon memo on the IC-to-manager ratio · Meta’s 2023 flattening · Gusto and Live Data Technologies figures on the shrinking middle-manager share.

Accountability Sinks

Why nobody in a large organization can be found responsible for anything, and why that is a feature for the people inside it. One chapter only — the project management manuscript covers this ground at length and two books shouldn’t both spend a hundred pages here.

EVIDENCE — Dan Davies, The Unaccountability Machine (2024) · David Graeber, Bullshit Jobs (2018) · Ringelmann and the social loafing literature · Parkinson’s law of triviality · cross-reference: Paleo Project Management, Ch. 6 and Ch. 9.

Outdoor Cats

The Greenville hang-up, told as evidence rather than memoir. Four managers on an ambush call, and one field technician who said the unsayable — and the only reason he could say it was that the bill for leaving was one he could personally cover. Indoor cat and outdoor cat becomes the working vocabulary here, framed as a trade rather than a hierarchy: one pays in stress about work, the other in stress about money.

SOURCE — Original. Pair deliberately with Ch. 1: the attorney and the network administrator are the same animal in different enclosures, and the difference between them isn’t character. It’s the photograph on the shelf.

Pearson’s Flip

State his argument fairly, credit it, and don’t re-litigate it. The risk-reward on working for yourself inverted; he told the individual, and the case is made. What follows from it for an employer is what this book is about.

EVIDENCE — Taylor Pearson, The End of Jobs — especially the content-versus-contextual-knowledge distinction · MJ DeMarco, Unscripted, and the CENTS framework, which leads with control · Naval Ravikant on permissionless leverage (code and media) · US Census Business Formation Statistics, elevated since 2020 · MBO Partners State of Independence · Blanchflower & Oswald: self-employed satisfaction runs higher despite worse pay and longer hours.

One Person, Whole Job

The fragmentation of Ch. 8, reversing. Task identity is returning as a matter of capability rather than philosophy — one competent person can now hold marketing, accounting, production, and distribution — and AI is the largest lever in that sequence, not a separate story.

EVIDENCE — nonemployer firm growth in Census data · the micro-preneur pattern across marketing, construction, film, software, and publishing · Paul Jarvis, Company of One · the honest counterweight: Autor and colleagues on superstar firms and market concentration — more solo operators and more concentration, with the middle hollowing.

The Fire Drill

2021–22 as a natural experiment rather than an anomaly: what a workforce does when the constraint briefly loosens, and what management did in response. The response is the more interesting half.

EVIDENCE — BLS JOLTS quits data at record levels across 2021–22, and the subsequent return toward baseline — report both · Nicholas Bloom et al., Nature (2024) — a randomized trial: hybrid cut attrition by roughly a third with no measurable performance loss · growth in employee monitoring software · Microsoft Work Trend Index on productivity paranoia · return-to-office mandates enforced by badge data.

Who Stays and Who Can Leave

The durable version of the exit argument, scoped honestly. Most people won’t leave — the lock in Ch. 5 is real and it holds. But mobility and capability correlate, which means an enclosure that retains by default is running a filter, and the filter selects against exactly the people it most needs.

EVIDENCE — Albert Hirschman, Exit, Voice, and Loyalty — specifically that exit by the most capable degrades the institution for everyone who stays · Reed Hastings, No Rules Rules (talent density) · Nassim Taleb, “How to Legally Own Another Person,” in Skin in the Game — engage it directly rather than letting a reviewer raise it · regretted-attrition data as the metric that actually matters, against gross turnover.

Band Size

The working unit stays small regardless of enterprise size. You scale by replicating small units, never by enlarging one — which turns out to be the answer to change management as well as to scale, since nobody has successfully flattened a big thing and several have successfully replicated small ones.

EVIDENCE — Dunbar’s number and the layers beneath it · W.L. Gore’s practice of splitting facilities before they outgrow coordination · Amazon’s two-pizza team and single-threaded owner · Brooks’s Law on communication overhead · cross-reference: Paleo Project Management, Ch. 20.

Custody

Give a person a whole, identifiable piece of work with a visible outcome and their name on it. Highest-yield change available, and it costs nothing. The spine of this book and of the project management one.

EVIDENCE — Hackman & Oldham on task identity — validated for fifty years, ignored for fifty years · Amabile & Kramer, The Progress Principle · Amy Wrzesniewski on job crafting and callings · original: what changes in a client engagement when the expert is emancipated rather than managed by committee.

Rank for the Hunt

Leadership should be real, total, and attached to a bounded task — then dissolve. The campus had a chief; what it didn’t have was eight of them stacked between you and the outcome. Modern organizations get this backwards twice: permanent rank for daily life, committees for the bounded work where someone actually needs to be in charge.

EVIDENCE — Toyota’s andon cord: anyone can stop the line, one person decides what changes · Pixar’s Braintrust: anyone flags, only the director decides · Kelly Johnson’s Skunk Works rules · Apollo 13’s tiger team as bounded, total, temporary authority · cross-reference: Paleo Project Management, Ch. 19, 27, 29.

The Kill Is Shared

The bonus-and-ownership question, answered with data. Ownership without control changes little; control without a share breeds cynicism. You need both, and there is real evidence on what each produces alone.

EVIDENCE — Rutgers Institute for the Study of Employee Ownership and Profit Sharing (Kruse, Blasi) on ESOP productivity, survival, and layoff resistance · Nucor’s production bonus and flat layer count · Mondragón · John Lewis Partnership · Kelly Johnson’s rule against tying pay to headcount supervised.

Trust Is Cheaper Than Surveillance

Monitoring costs money to install, costs output to run, and reliably produces the behavior it was meant to catch. Make the cost side explicit — most executives have never seen it totaled.

EVIDENCE — Ethan Bernstein’s transparency paradox (2012) — observed workers performed worse · Amy Edmondson, The Fearless Organization, and Project Aristotle on psychological safety · Deci’s undermining effect · Fried & Hansson on meeting cost arithmetic.

Building the Glade

Ch. 2 without a unicorn’s budget. Which features of that campus actually did the work, which were expensive theater, and what the cheap versions are — daylight and sightlines and moving water and plants cost far less than massages and free lunch, and the evidence says they matter more. Includes the open-plan correction: the camp is a small band with acoustic privacy, not four hundred people in a room.

EVIDENCE — Bernstein & Turban (2018) · DeMarco & Lister, Peopleware, on noise and interruption · Roger Ulrich’s hospital window recovery study · Kaplan & Kaplan, attention restoration theory · Jay Appleton, prospect-refuge · E.O. Wilson, biophilia · Terrapin Bright Green, 14 Patterns of Biophilic Design · the honest note that perks are the cheapest thing to copy and the least load-bearing, which is why so many companies copied them and got nothing.

Metering Innovation

Make rate of improvement a tracked number, for the product and for the process that makes it. Anything unmeasured loses to anything measured — which is precisely how bureaucracy wins by default, and how the invisible waste of Ch. 7 becomes visible.

EVIDENCE — Hamel & Zanini’s bureaucratic mass index as a usable instrument · Bayer’s rulebook reduction as deletion-as-objective · Goldratt’s Theory of Constraints and the Five Focusing Steps · the question-delete-simplify-accelerate-automate sequence, strictly in that order · cross-reference: Paleo Project Management, Ch. 14 and Ch. 24.

Manage on the Hire

Front-load diligence into selection instead of back-loading it into surveillance. The least-managed teams produce the best work because the judgment happened before anyone started.

EVIDENCE — Buckingham & Coffman, First, Break All the Rules — the manager accounts for most of the variance in team engagement · Hastings on talent density · Laszlo Bock, Work Rules!, on Google’s hiring data and the weakness of unstructured interviews · Schmidt & Hunter on structured interview validity.

The Good Jobs Proof

The strongest evidence available, from the least glamorous industry imaginable. If it works on retail margins, nobody gets to call it a software-company luxury — and the 2023 book matters more than the 2014 one because it documents conversions rather than companies that were always good.

EVIDENCE — Zeynep Ton (MIT), The Good Jobs Strategy (2014) and The Case for Good Jobs (2023) · her cases: Costco, QuikTrip, Mercadona, Trader Joe’s · Gallup Q12 meta-analyses comparing business units within the same firm · Bloom’s randomized trials as the cleanest causal evidence in the field · concede the reverse-causality problem before a reviewer raises it: profitable firms can afford good jobs.

Where It Broke

The credibility chapter, and the one with the uncomfortable finding: nearly every success was either built this way from the start or converted by an owner with absolute authority. Transformations inside firms with dispersed power mostly fail or get quietly reversed — which is also the reason Ch. 2’s campus was a startup with a founder still in the chair.

EVIDENCE — Zappos’s Holacracy adoption, the attrition that followed, and the walk-back · Basecamp’s 2021 policy change and departures · Semco and Haier as conversions that required unchallengeable owner authority · Gore, Nucor, Morning Star, Buurtzorg as founded-that-way cases · the honest epilogue on the campus itself: what happened to 20% time as the company scaled · note: the “70% of change programs fail” figure is folklore with no study behind it — don’t use it.

Can Flat Go Global

Yes — but never by flattening one large thing. By replicating small ones. Probably the most original argument in the book, and the answer to Ch. 25’s problem: the unit of change is the team, which is why a pilot works where a transformation doesn’t.

EVIDENCE — Haier’s RenDanHeYi — thousands of microenterprises inside an appliance multinational · W.L. Gore’s lattice and plant-splitting across decades · Buurtzorg’s small autonomous nursing teams in a regulated system, with published cost and outcome results · Hamel & Zanini, Humanocracy · Corporate Rebels / Joost Minnaar for further documented cases.

What I Ran

Your own numbers, including the ones that failed. The only chapter nobody else could write, and what separates this from a synthesis of other people’s research.

SOURCE — FlashPointLabs, GrowMyPractice.legal, RemodelCarpenters.com: voluntary turnover against benchmark, revenue per head, loaded replacement cost, contractor spend against equivalent headcount. Plus siloz.com and the earlier ventures as the failure set. Instrument these before drafting — the numbers can’t be reconstructed later.

Conclusion: Change the Arithmetic

Return to the two rooms. The man in the box isn’t weak and isn’t wrong; he’s doing the math anyone would do with those inputs. The campus proves the other room is buildable — and Part IV proves it’s buildable without a unicorn’s balance sheet. A system that sort of works is the hardest kind to fix and the most expensive kind to leave alone.

NOTE — End on an instruction, not a manifesto. End on him, not on you. And be honest that most companies won’t do this — some will decompose, others will lose their best people to firms that got it right from the start. A more believable ending than mass conversion, and a better one.

The Manifest

The Animal & the Box

How the Box Got Built

The Conditions Changed

Build for the Animal

Does It Pay