Subscribe now

The Advantage That Learned to Say No

For most of the twentieth century, General Motors ran on a simple, ruthless piece of financial logic: fund what pays back fastest, and let everything else compete for what’s left. It built an empire. Eventually, it also built a bankruptcy.

The Machine That Beat Ford

In the 1920s, Alfred Sloan paired decentralized, market-facing divisions with tight, centralized financial control — a structure built to coordinate five brands into a single price ladder, from Chevrolet to Cadillac. It worked, propelling GM into decades of dominance as the world’s largest automaker. The discipline wasn’t incidental to the success. It was the mechanism of the success.

That discipline needed a home, and it found one: a New York finance and treasury organization that became, in the words of one industry account, “an elite within the company, a breeding ground for CEOs.” Roger Smith, Jack Smith, and Rick Wagoner — three consecutive chairmen spanning nearly half a century — all rose through that same Treasurer’s office. GM hadn’t just built a strong finance function. It had built a company where finance was, structurally, the top.

Every Decision, Run Through One Filter

Once that pipeline hardened, something quieter happened: finance’s judgment stopped competing with manufacturing, marketing, and design — it started overruling them. Decisions by GM’s finance staff “took precedence, often overruling initiatives” from the divisions that actually made the cars.

You can watch the pattern repeat across three decades of unrelated-looking decisions. When Cadillac rushed a rebadged Chevrolet Cavalier to market as the 1982 Cimarron, Cadillac’s own president objected that there wasn’t time to properly differentiate the car — and was overruled anyway. Cost logic won; brand logic lost. It became the industry’s favorite example of badge engineering gone wrong, but it was never really about badges. It was about which department’s math got the final word.

The Trade-Off Nobody Voted On

Sharing platforms across Chevrolet, Pontiac, Oldsmobile, Buick, and Cadillac cut costs beautifully — and quietly dissolved the price ladder that had justified paying more for a Buick than a Chevrolet in the first place. By the late 1990s, Oldsmobile and Buick models sat within a few hundred dollars of each other, and nobody inside or outside the company could say why.

The same filter reached further than pricing. When GM’s EV1 drew a genuinely enthusiastic customer response in the late 1990s, the company recalled and destroyed most of the leased units anyway, on the grounds that electric vehicles were “an unprofitable niche.” Saturn, launched with real early success, lost its internal champion the same year its first car rolled off the line — and was immediately starved of capital once company-wide losses put every division in competition for what was left. Between 1996 and 2000, GM returned more than $20 billion to shareholders through buybacks and dividends, capital that was, by definition, no longer available for any of this.

None of these were reckless decisions. Each one, taken alone, was exactly what a disciplined finance organization should choose: fund the fastest payback, cut the underperformer, return excess capital. That was the whole point of the machine Sloan built.

When the Storm Came

By 2008, the mechanism had done its work. Brand distinctiveness was gone. The technology hedge was gone. The patient bet was gone. What remained was a company more dependent than ever on its highest-margin, fastest-payback segment — large trucks and SUVs — just as a financial crisis and an oil-price spike hit that exact segment hardest. There was no cushion left, because the cushion had never scored well enough to survive.

The Decoded Insight

The discipline that builds an organization’s early advantage is rarely the thing that kills it directly. What kills it is leaving that discipline unchallenged long enough that it becomes the only lens through which every future decision gets made — because the qualities a durable advantage actually requires, patience, distinctiveness, hedges against an uncertain future, are precisely the things that discipline’s own metrics were never built to reward.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Deal That Never Came

In October 1999, WorldCom announced a $129 billion agreement to acquire Sprint — framed not as expansion but as inevitability, the move that would make WorldCom the definitive leader of American telecommunications, ahead of AT&T itself. Eight months later, the Department of Justice sued to block it. By mid-July 2000, the companies had mutually terminated the agreement.

Within weeks, according to the SEC, WorldCom’s senior management began instructing accountants to overstate earnings — an overstatement that would run for at least seven consecutive fiscal quarters. It’s tempting to read this as coincidence, a company already inclined to cheat simply choosing this moment. The evidence suggests something more specific: the fraud’s documented start lines up almost exactly with the month WorldCom’s next source of growth disappeared.

The Machine That Ran on Acquisitions

WorldCom’s rise was never a story about operational excellence in long-distance telephony. It was a story about a loop: a rising stock price made each acquisition cheaper to finance with more stock, each acquisition added revenue and scale, and that growth justified the rising price that funded the next acquisition. Run long enough, this loop turned a Mississippi discount reseller into one of the most heavily valued companies in the country. It was not irrational — it worked, repeatedly, for over a decade.

But the loop carried a quiet dependency: the company never had to prove it could grow without acquiring something, because it never stopped acquiring long enough to find out. Sprint was simply the next leg of a very long relay.

When the Baton Wasn’t There

WorldCom’s own board came to recognize, after the Sprint deal collapsed, that Ebbers had no coherent strategy to replace it. That’s a striking admission — not evidence of secrecy, but of an organization watching the gap open in real time and not knowing how to close it through ordinary means.

The market, meanwhile, didn’t know the baton had been dropped. WorldCom’s growth expectation — priced into the stock, built into analyst models — didn’t shrink to match the deal’s absence. Nobody had agreed to revise it downward. A number the business could no longer legitimately produce was still, as far as the market was concerned, owed.

Manufacturing What the Business Could No Longer Grow

This is the shape the fraud actually took. CFO Scott Sullivan later explained the capitalized line costs by invoking the accounting concept of matching costs to revenue — telling investigators the network capacity WorldCom had built out was still waiting for usage growth that had simply never arrived. It was, in its way, an honest description of the underlying problem, offered as a defense of a dishonest fix. KPMG’s later verdict to the Audit Committee was blunt: nothing in GAAP supported the entries.

The accounting fraud wasn’t the only mechanism protecting the story. In the same window, WorldCom’s board authorized a running series of loans and guarantees — roughly $400 million in total — to keep Ebbers from having to sell his stock on the open market to cover margin calls. Two parts of the same organization, on two different fronts, were both working to keep the story from visibly cracking.

The Decoded Insight

It cracked anyway — not through a market correction, but through an internal auditor’s quiet, after-hours review of a handful of unexplained entries. What she found wasn’t a company that had lied about its business. It was a company still paying, in every ledger and every boardroom decision, for a growth number that a blocked merger had made impossible to earn honestly.

A growth story that depends on a repeatable external event — the next deal, the next round, the next approval — carries a specific vulnerability that has nothing to do with the character of the people running it. It fails not when the event disappears, but when no one is willing to be the one who tells the market it disappeared. In that silence, the gap between promise and business gets filled with something. At WorldCom, for nineteen months, it was filled with numbers instead of news.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Turnaround That Never Got to Finish

For most of a decade, Borders’ board had a reliable answer to bad news: replace the strategy. It felt like accountability. It was actually the opposite of what the company needed.

Why It Worked

Robert DiRomualdo ran Borders for roughly ten straight years, from 1989 to 1999. He built one thing, consistently: superstore expansion powered by a proprietary inventory system so effective a publisher once called it “the envy of the industry.” That continuity produced the 1995 IPO and turned a college-town bookstore into a national chain. Nothing clever about the mechanism — just one person, one strategy, ten years to let it compound.

The Hidden Trade-Off

What Borders never built was a governance habit of extending strategy. It had only ever known one leader’s tenure. When DiRomualdo left, the board had no principle to fall back on — only a template: when performance dips, bring in someone new to undo what came before.

The Trap

That template ran, almost mechanically, for the next twelve years. Greg Josefowicz, hired from supermarket retailer Jewel-Osco, spent seven years building category management and an international expansion into the UK, Australia, and Singapore. His successor, George Jones, arrived with an explicit mandate to reverse both: he killed category management, sold off most of the international stores, and clawed the company’s website back from Amazon, where it had been outsourced since 2001. Jones got two and a half years before he was fired. His replacement, Ron Marshall, a private-equity turnaround specialist with no book-industry background, spent roughly a year redirecting the company again — this time toward pure cost-cutting — before resigning.

Each hire’s job was implicitly to prove the last person wrong. None stayed long enough to prove themselves right. The website, the international stores, the cost structure — each strategy was assembled, then dismantled, before its results had time to show up in the numbers.

The Breaking Point

The one capability Borders needed most in its final decade was digital retail — a slow, cumulative competency that only compounds with years of consistent investment. It arrived exactly when CEO tenures had shrunk to eighteen months and under. Borders didn’t attempt a serious e-book push, through a partnership with Kobo, until 2010 — the same year its chairman and CEO both changed again. By then Amazon’s Kindle had a three-year head start, and Barnes & Noble’s Nook had a year on the market.

Barnes & Noble is the clean comparison, because it faced identical disruption inside an identical format. In 2000, its annual report committed to building its own e-commerce site even at the risk of undercutting its own stores — and it held that position long enough to launch the Nook in 2009. Same industry, same pressure. The difference wasn’t foresight; it was that one company let a strategy run, and the other kept restarting it.

The Decoded Insight

Reversal feels like discipline. A board that fires an underperforming leader and installs someone with a different plan looks decisive, willing to change course. But reversal and correction are not the same act. Correction adjusts a strategy while keeping its direction intact long enough to see whether the adjustment worked. Reversal discards the direction itself — and with it, whatever time had already been invested toward making it pay off.

An organization that treats every setback as grounds for a new strategy never actually tests any strategy. It just keeps resetting the clock, indefinitely, on the one thing — digital retail, in Borders’ case — that could only ever have been won by letting the clock run.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Layer Motorola Never Believed In

For six years, Motorola kept changing its mind about the same question, and never noticed it was the question that mattered most.

The Plumbing Theory

In February 2003, standing in front of an industry starting to argue over which mobile operating system would win, a Motorola executive said something remarkably candid: the OS itself wasn’t strategic. What mattered was everything built on top of it. The company was committing its phone lineup to Linux and Java — but the commitment came wrapped in a belief that made the choice feel almost irrelevant. If the operating system was just plumbing, it didn’t much matter which pipes you used.

That belief wasn’t rhetorical. Just months after declaring itself “full square behind Linux/Java,” Motorola was quietly building a Microsoft-powered smartphone on the side. Not a betrayal of strategy — a natural extension of one. If the OS is replaceable, why not keep a second one warm?

Why It Worked

The belief had evidence behind it, and the evidence was enormous. The RAZR V3, launched in 2004, was a design triumph — a 13.9mm aluminum clamshell that made every other phone on the market look thick and plastic. Motorola sold it for four years with barely a hardware revision. It moved 130 million units. Nobody was choosing a RAZR for its software. They were choosing it because it was the best-made object in their pocket. Hardware was where Motorola won, and RAZR proved it at a scale few consumer products ever reach.

The Hidden Trade-Off

But a belief validated that thoroughly stops behaving like a belief. It starts behaving like a fact about how the industry works. A company that “knows” the OS doesn’t matter has no reason to protect any single software investment from being abandoned the moment something more urgent shows up. So none of Motorola’s platform bets were ever meant to be permanent. They were hedges — reasonable, even prudent, from inside a worldview where hardware was the real product and software was interchangeable underneath it.

The Trap

Between 2006 and 2008, the company’s global handset share fell from roughly 22% to single digits, even as Motorola kept iterating on hardware and kept shuffling operating systems underneath it. In 2008, it tried again with the Motorola Q, a Windows Mobile device aimed squarely at the BlackBerry crowd. It flopped even at $99. The company had built plenty of good phones. It had never built a platform, because it had never let one sit still long enough to become one. Each restart reset the clock on the only kind of advantage that compounds: accumulated software depth.

The Breaking Point

In January 2009, Motorola shut down MOTOMAGX — the Linux/Java platform it had been investing in since 2003 — and pivoted to Android. Sanjay Jha, then co-running the mobile business, put it plainly: the company needed to get better at “execution of our software strategy.” Six years, three platform strategies, and the admission still described the problem as execution, not belief. A study the year before had already found that 24% of new iPhone buyers in the U.S. were switching from a RAZR. Motorola’s most loyal, hardware-loving customers were the first to leave the moment the product that mattered was no longer the object in their hand but the ecosystem behind the screen.

The Decoded Insight

Motorola didn’t fail from a shortage of conviction. It failed from a surplus of the wrong kind — a belief about where value lived that had been proven so completely, on such a massive scale, that it no longer needed defending. That’s what makes a company’s greatest strength quietly dangerous: not that it stops being true, but that it stops being questioned, right up until the ground it was built on shifts underneath it.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Rescue Deal Nissan Never Renegotiated

In November 2018, the man credited with saving Nissan from near-bankruptcy was arrested at Haneda Airport by Tokyo prosecutors, not removed by a vote of his own board. Inside Nissan’s Tokyo headquarters, several outlets reported applause when the news broke. The man who had rescued the company was being cheered out of it — through a criminal referral, not a boardroom decision. That gap between the two channels is the actual story.

The terms that eventually produced this outcome were set in 1999, when Nissan was in genuine crisis. Renault agreed to rescue it in exchange for a 43.4% voting stake in Nissan; Nissan received a 15% non-voting stake in Renault in return. That asymmetry wasn’t an insult — it was the price of survival, and it worked exactly as intended. Carlos Ghosn arrived with enough real backing to force through reforms a purely domestic board might never have managed on its own. The imbalance was the mechanism that made the turnaround possible.

What the 1999 terms didn’t include was an expiration date. As Nissan grew into the larger, more profitable half of the alliance, its formal voice in governing it never grew to match. Success accrued to the company; institutional power did not follow. And when the balance of power did come up for discussion, it moved in the wrong direction. In 2014, a new French law doubled the voting rights of long-term shareholders, and the French state used its Renault stake to extend its leverage over the alliance further — deepening the very asymmetry Nissan had inherited from its own rescue.

Nissan’s leadership was not passive about this. In September 2015, the company formally proposed rebalancing the cross-shareholding to something closer to 25–35%, with equal decision-making weight. It went nowhere. This is the detail that matters most: Nissan tried the legitimate route years before the crisis, and was rebuffed. Each failed attempt at correction left the underlying disagreement with one fewer legitimate outlet — and one more reason to eventually look for a channel outside the boardroom. By early 2018, when Ghosn began exploring a full three-way merger of Renault, Nissan, and Mitsubishi, Nissan’s executives faced a chairman who held the top post at both companies, and a governance structure that gave them no formal vote to stop him if they disagreed.

What followed was not a governance solution but a workaround. The internal investigation that preceded Ghosn’s arrest was led by Hari Nada, a senior Nissan executive who was later investigated himself, over a separate compensation conflict of interest. Ravinder Passi, the general counsel who raised concerns that Nada’s own involvement compromised Nissan’s ability to defend itself, was demoted and ultimately pushed out — and later sued the company for wrongful termination. The channel that replaced the boardroom had no independent check on it either. It took another five years of public crisis before the underlying imbalance was actually corrected: in 2023, Renault and Nissan finally equalized their cross-shareholding at 15% each, moving the disputed portion of Renault’s stake into a neutrally-voted trust. The fix that had been proposed in 2015, and was structurally available all along, arrived only after the arrangement had already produced the very rupture it was meant to prevent.

Governance terms built for an emergency don’t expire on their own. If they aren’t revisited as the underlying balance of power shifts, the imbalance doesn’t quietly correct itself — it becomes a standing source of conflict that outlives its own justification. Dissent inside that structure doesn’t disappear when it has nowhere legitimate to go. It waits for the first available crack, and it exits through that crack however destructively it has to.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Machine That Sold Calm

For twenty years, General Electric delivered something almost no industrial company could: predictability. Quarter after quarter, earnings landed where Wall Street expected. Investors paid a premium for that certainty. What they didn’t know was that the certainty was manufactured.

The Engine Behind the Calm

GE Capital was the mechanism. Using GE’s AAA credit rating, it borrowed cheaply and deployed capital across leasing, lending, and insurance, smoothing the industrial side’s lumpier results into one reliable number. By the mid-to-late 2000s, GE Capital-related businesses generated roughly $60–70 billion in annual revenue — about 40% of GE’s total — on nearly $500 billion in assets. GE became the standard other conglomerates were measured against, precisely because it never seemed to surprise anyone.

What the Calm Was Costing

The consistency wasn’t organic. Between 1997 and 2001, GE underfunded reinsurance reserves by an estimated $9.4 billion. Strip that out, and profit growth over the final five Welch years falls from a reported 90.2% to roughly 5.6%. This was legal and audited, unremarkable by the standards of the era. But the “predictability” investors priced in was an output of the machine, not a description of the business. Every division had effectively signed up to keep the story consistent.

The Trap

Once “beats every quarter” became GE’s identity, the organization couldn’t announce that the underlying businesses were more volatile than reported without repricing the entire company. So managing the picture rather than reporting it became institutional habit — one that outlived the executive who started it. In February 2009, GE cut its dividend for the first time since 1938, from $0.31 to $0.10 a share, conserving roughly $9 billion a year: the first public admission that cash flows could no longer support what the smoothed numbers had implied.

Rather than shrink its ambitions, GE tried to out-earn the problem. In 2015 it closed its largest-ever industrial acquisition, buying Alstom’s power business for roughly $10.6 billion, projecting steady EPS gains through 2018 — a bet that one more big move could restore the growth story the financial engine used to supply. By 2017, with gas-turbine demand collapsing, GE Power’s profits had fallen 45%, and GE cut 12,000 jobs from the unit while halving its dividend again.

The Breaking Point, Twice

The 2008 credit freeze made the original engineering impossible to fund — the first breaking point. The more revealing one came a decade later, in a business that had nothing to do with GE Capital. In January 2018, GE disclosed a $6.2 billion insurance charge and forecast $15 billion more in reserves needed over seven years, for a long-term-care business it had told investors was already de-risked. Former employees later testified that senior executives had changed reserve figures and methodology without supporting evidence. GE Capital was already being dismantled by then. It was the same habit, resurfacing in whichever unit was next asked to keep the numbers calm.

The Decoded Insight

An organization that learns to sell consistency instead of reporting results cannot simply stop selling it — admitting the truth costs more, in the moment, than continuing to manage the picture. So the practice doesn’t end when the division that invented it is wound down. It migrates to the next set of books that can absorb the smoothing, until the gap between story and balance sheet becomes too large for any single division to carry. What looks, in hindsight, like a string of unrelated failures — a financial crisis, a bad acquisition, an insurance shortfall — is usually one habit, wearing different clothes.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The App That Ran Out of Time

Clubhouse once looked like the fastest-growing consumer product in Silicon Valley history. Then it lost most of its download volume within about a year — not because a rival built a better app, but because the world went back to work.

The standard account blames the clones. Twitter, Meta, Spotify, and LinkedIn all rushed out live-audio features within months of Clubhouse’s breakout, and the timeline seems to confirm the story: competitors arrived, Clubhouse shrank. But the sharpest collapse in Clubhouse’s install numbers — a 73 percent drop between February and March 2021 — happened at the very peak of its hype, before most of those clones had scaled to compete with anything. Something else was already breaking.

A Format Built on Borrowed Time

Clubhouse launched in March 2020 with a single, uncompromising design choice: every conversation was live, synchronous, and unrecordable. There was no feed to scroll later, no episode to catch up on, no way to participate except by showing up in real time. During lockdown, that constraint was invisible, because real-time attention was suddenly the one resource nobody was short of. People had nowhere else to be. A format that demanded pure presence felt less like a limitation and more like intimacy.

The Trade That Looked Like a Feature

What made the product feel special was also what made it fragile. By refusing to let users record or drop into a conversation asynchronously, Clubhouse didn’t just favor a format — it made a bet that its users’ spare, uncommitted time would remain abundant indefinitely. Early growth seemed to confirm the bet was working, so the company reinforced it rather than hedging it: nothing to soften the requirement that you be there, live, or get nothing at all.

The Trap of Validated Success

Because the live-only design was the thing driving adoption, there was no internal signal telling the team to change it. Growth curves don’t distinguish between a habit that’s durable and a habit that’s borrowed from a temporary circumstance. Asynchronous features — Replays, Clips — didn’t arrive until October 2021, well after the user base had already crested and begun its decline. By then, the daily ritual of dropping into a room had already lost its place in people’s schedules to something else.

The Breaking Point

That something else was simply normal life resuming. As offices reopened and calendars refilled, the resource Clubhouse’s format required — unclaimed, synchronous stretches of time — reverted to scarce. A comparison with Twitter Spaces is instructive here: same audio format, same historical moment, same pandemic tailwind, but embedded inside an app people already opened for other reasons, so a Space could be joined for two minutes between other things rather than requiring a dedicated block of attention. One retrospective estimate put six-month retention near 8 percent for Clubhouse against roughly 67 percent for Spaces — a gap suggestive enough, even allowing for the imprecision of any single estimate, that feature parity alone can’t explain it.

By April 2023, when Clubhouse’s founders cut more than half the staff, they described the problem in almost exactly these terms: it had become harder for people to find their friends on the app and to fit long conversations into their daily lives. Tellingly, they said the company still had years of runway — this wasn’t a cash crisis. It was a product whose core requirement had quietly stopped being available.

Clubhouse didn’t lose a competitive war. It ran on a resource — other people’s uncommitted real time — that was never going to stay abundant. The clones arrived, but they arrived at a house whose foundation had already been built on borrowed time, and no update, no matter how fast, restores a condition that was only ever temporary to begin with.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Signal That Never Reached Anyone

Lehman Brothers had good risk managers. It had a warning system that worked exactly as designed. In June 2005, Michael Gelband, the head of global fixed income, stood in front of his colleagues and told them the real estate market could collapse within two or three years. The system caught the problem early. What happened next is the story.

Why It Worked

Lehman’s growth engine ran on short-term repo financing — borrowing against securities, rolling it over constantly, scaling into mortgage origination and securitization without the drag of raising permanent capital. Between 2000 and 2007, this let assets more than triple to $691 billion. It was a fast, efficient way to compound growth, and for years it worked exactly as intended. But a model built on continuous short-term borrowing has one silent precondition: the lenders have to keep believing you’re fine. Every day.

The Hidden Trade-Off

That precondition quietly became the organization’s real priority. In February 2007, when the same fixed-income chief who had warned of a “train wreck” repeated his concerns, the firm’s COO took him to lunch and told him he was “too conservative” — that he was costing the firm deals. The trade-off wasn’t a single bad call; it was a standing preference, reinforced every time someone chose growth over the discomfort of listening to dissent.

The Trap

By September 2007, the chief risk officer, Madelyn Antoncic, had repeatedly objected to routine waivers of the firm’s own risk limits. She was removed from the executive committee. Around the same time, the finance function began leaning harder on an accounting maneuver known internally as Repo 105 — quarter-end transactions structured so that borrowed money could be booked as a sale, temporarily erasing the debt from the balance sheet. At the end of 2007’s fourth quarter, $38.6 billion vanished this way. By the second quarter of 2008, it was $50.4 billion, timed precisely to land before each reporting date. The two systems that could have forced a correction — a risk officer with standing to say no, and a balance sheet lenders could actually trust — were disabled by the same impulse, for the same reason. Reported leverage at that point stood at 12.1x. The real number, without the accounting device, was 13.9x.

The Breaking Point

Neither channel had to fail for Lehman to be in danger. Both had to fail for the danger to become invisible to everyone positioned to act on it. As scrutiny of Wall Street balance sheets intensified in 2008, what surfaced at Lehman wasn’t a risk that had just appeared. It was a risk that had been accumulating, undisclosed and unchallenged, for two years. There was no gradual unwind available anymore, because gradual unwinds require someone inside the building or outside the door catching the problem while it’s still small. Both of those people had already been dealt with.

The Decoded Insight

Even after collapse, Fuld maintained he hadn’t understood what Repo 105 was doing to the balance sheet — despite records showing the relevant material had reached him months earlier. That denial fits the pattern more than it breaks it: an organization that spends years training itself not to hear bad news eventually produces leaders who genuinely can’t. The mechanism here isn’t recklessness. It’s that any entity whose survival depends on someone else’s continuous confidence will always face a choice between protecting that confidence and listening to what threatens it — and choosing the former doesn’t make the risk disappear. It just guarantees that when the reckoning comes, it comes all at once, to people who never saw it building.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Discount That Wasn’t a Discount

For years, JCPenney’s customers had a ritual. Circulars arrived, coupons got clipped, prices dropped in predictable cycles, and shoppers came away feeling like they’d won something. In February 2012, the company decided this ritual was inefficient — a maze of “fake prices” standing between customers and honest value — and eliminated coupons, sales, and promotions in a single stroke, replacing them with flat, everyday pricing.

By the end of the year, JCPenney had lost $985 million.

The System That Wasn’t Broken

The decision looked reasonable on its face. High-low pricing is, mechanically, a game of inflated list prices followed by discounts that make customers feel clever for waiting. Collapsing that into one honest number isn’t obviously a mistake — it looks like removing friction. And for years, the friction-filled system had worked: customers came back again and again, not because the underlying prices were unusually low, but because the hunt for a deal had become part of how they shopped.

That’s the detail easy to miss in hindsight: the coupon wasn’t decorating the transaction. For a meaningful share of JCPenney’s customers, it was the transaction — the reason to walk in, and the payoff for walking in.

The Trade-Off Nobody Priced In

Leadership treated the promotional system as overhead — a cost center to streamline, not infrastructure to protect. Nothing in the rollout suggests anyone asked what would happen if the ritual itself, rather than the prices behind it, was what kept customers engaged. The company’s own president later admitted as much: coupons, he said, “drove traffic and conversions,” and management “did not realize how deep some of the customers were into this.”

That single sentence carries the whole trade-off. Success had been quietly built not on price levels but on a reward loop — one the company had never recognized as the actual asset driving the business.

The Trap

Once the ritual was gone, the collapse wasn’t evenly distributed. First-quarter comparable sales fell 18.9%, but the damage inside that number was telling: weekend traffic — the slot most associated with bargain hunting — fell 12% year-over-year, nearly double the 6% weekday decline. The customers most attached to the ritual left first, and left fastest.

By year’s end, sales had fallen 24.8% to $12.985 billion, the company’s lowest revenue since 1987. A real, gradual decline had been building for years — revenue had already slid from $20.2 billion in 2007 to $17.8 billion in 2011 — but nothing in that slower slope explains a single-year drop of this magnitude. The gap between the two is the mechanism’s fingerprint.

The Breaking Point

JCPenney tried to reverse course, bringing back coupons roughly seven months later, then clearance racks soon after. But by September, the problem wasn’t pricing anymore — it was trust and habit, and habits don’t reboot on command. The CEO who had championed the strategy was removed the following April, seventeen months after he’d started.

The Decoded Insight

What collapsed at JCPenney wasn’t a pricing scheme. It was the quiet discovery that a ritual built to drive engagement had become indistinguishable from the product itself — so removing it, however rational the spreadsheet made it look, wasn’t simplification. It was deletion. Any business whose relationship with customers runs on a reward loop — a discount cadence, a points system, a game of finding the deal — carries the same risk: the loop can quietly become the product long before anyone notices, and by the time someone calls it “friction,” the friction is the business.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The contradiction in this story is that the man who trusted no one eventually had almost no one left to trust him back.

By the mid-1960s, one person stood between Howard Hughes and nearly everything else in the world. Robert Maheu ran the Las Vegas operation, negotiated the casino purchases, and managed the daily business of one of the richest men alive — while Hughes himself, holed up behind blanketed windows and buffer rooms rented on either side of his own, went weeks or months without being seen by his own staff. This was not a sudden collapse. It was the end point of a strategy Hughes had built, retreat by retreat, over three decades — a strategy that had once made him formidable.

Hughes’s early success rested on a genuine strength: he didn’t need anyone’s permission. He piloted his own experimental aircraft, financed his own films, and negotiated his own business deals, often over the objections of people who thought he should delegate more. In 1947, defending himself before a Senate committee investigating his wartime aircraft contracts, he put the philosophy plainly: he had gotten into trouble, he told the senators, by “doing too much of it with my own two hands.” Months earlier, that instinct had nearly killed him — he personally piloted an experimental reconnaissance plane past a documented mechanical malfunction rather than hand off control or cut the flight short, and crashed it into a Beverly Hills neighborhood, barely surviving.

What followed each of these crises is the more revealing pattern. Rather than building in more outside checks, Hughes built in fewer. After the 1946 crash and the 1947 hearings, he moved toward deeper seclusion. When TWA became entangled in a multi-year legal dispute over control of the airline in the 1960s, Hughes declined to appear personally in court, even as the outcome determined whether he’d keep the company he’d run for two decades. He didn’t. In 1966, roughly 6.5 million of his TWA shares were sold off in twenty minutes, netting him over $500 million — capital that funded what looked, on the surface, like reinvention: a new chapter buying up Las Vegas real estate.

But it ran through the same closed structure. The Desert Inn, first of seven Nevada casinos he’d eventually own, wasn’t a deliberated entry into gaming — it followed after Hughes refused to vacate a suite he’d overstayed, and an aide suggested, half-joking, that he simply buy the hotel instead. Neither Hughes nor his top executives had run a casino before. What followed were years of suspected mismanagement, filtered through Maheu, with no internal mechanism positioned to flag or correct it.

Each retreat made sense on its own terms — privacy avoided intrusion, skipping court avoided an ordeal, one trusted aide simplified an increasingly complicated life. Stacked over twenty years, though, they produced an information environment with exactly one channel in and out, and no one inside that channel had the standing to intervene as Hughes’s own condition deteriorated. When he died in 1976, en route to a Houston hospital, his body had wasted to roughly ninety pounds, hair and nails badly overgrown, hypodermic needles broken off beneath his skin. The FBI needed his fingerprints to confirm it was him. One examining doctor gave the cause as neglect.

The familiar version of this story treats Hughes’s ending as a mind that simply broke. The more useful version is structural: isolation, narrowed far enough, stops just protecting the person inside it and starts removing anyone capable of noticing, or acting, when something goes wrong. The mechanism isn’t unique to eccentric billionaires. Any leader who filters the world down to a single gatekeeper, however loyal, is trading correction for comfort — and may not know the trade was made until no one is left to make it.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Advantage That Became a Liability

In 1991, Sega of America held something rare: the authority to make pricing and marketing calls, and the market knowledge to make them well. Tom Kalinske’s team ran an aggressive campaign around the Genesis, and within two years Sega’s US market share went from roughly 10 percent to roughly 60 percent. The company that had trailed Nintendo for a decade was suddenly setting the terms of the fight.

Nobody in Tokyo needed to approve the next move. That was the point. The people closest to American retailers, pricing psychology, and consumers were the people deciding what to do about them.

It would not stay that way.

As Sega’s next hardware generation approached, the stakes of each decision grew, and so did headquarters’ appetite for control. Sega of America had negotiated a joint hardware venture with Sony — cost-sharing, engineering collaboration, a business case Sony’s own Nobuyuki Idei agreed was sound. Sega of Japan killed it, doubting Sony’s seriousness. Around the same time, a dispute over the Saturn’s graphics hardware played out the same way: Sega of America wanted 3dfx chips; Tokyo chose PowerVR instead. Individually, these looked like ordinary corporate disagreements. What they shared was a pattern: as each decision’s cost of failure rose, the party with less exposure to the market took the wheel.

The pattern reached its full expression on May 11, 1995, on stage at the first E3. Kalinske had argued, vehemently, against launching the Saturn five months early. Sega of Japan overruled him anyway. He got up and announced the console was already on sale.

The people who weren’t ready found out in real time. KB Toys, left out of the limited retail supply, was so offended it stopped carrying new Sega hardware, permanently. Sony’s answer came from the same stage: Steve Race walked to the microphone and said one word — “$299” — erasing Saturn’s pricing advantage before the room had finished processing the surprise. A single centralized decision, made without the information the regional team had been trying to supply, cost Sega a major retailer and its pricing position on the same afternoon.

By mid-1994, in the middle of this same stretch, Sega’s US market share had already slid from 65 percent to 46 percent. In 1996, Kalinske left the company, naming the overruled launch as one of the central reasons.

None of this required anyone to be careless. Every override was a defensible exercise of authority by people trying to protect a growing, expensive bet. That’s what made it hard to see happening: centralizing control over the biggest decisions feels like discipline, not risk. But the information that made Sega’s earlier calls right didn’t move to Tokyo along with the authority to make them. It stayed with the team that had just been overruled.

When PlayStation 2’s momentum arrived a few years later, it met a company whose retailer relationships and market credibility had already been hollowed out, one high-stakes override at a time.

This is not a story about culture clash, or a company that stopped innovating. It’s a story about what happens, in any organization, when the right to decide and the information needed to decide well start traveling in opposite directions — and the gap between them opens widest exactly when the decisions matter most.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Confidence That Couldn’t Hear Itself Corrected

The contradiction in this story is that the instinct which built the Macintosh is the same one that got its creator removed from managing it.

In the early 1980s, Steve Jobs redirected Apple’s next major product away from a modest, low-cost machine and toward something more ambitious: a graphical interface adapted from ideas Xerox had shelved. The engineer who’d designed the simpler machine objected, formally and in writing, to Apple’s president and chairman. Jobs pushed forward anyway. A year later, he insisted on running a television ad the board itself was skeptical of. Both bets paid off. The Macintosh launch became one of the most studied product moments in business history.

That’s the part of the story people remember. What gets flattened is what happened next.

Being right against consensus tends to teach a lesson, whether or not anyone means to learn it: whose judgment to trust the next time someone disagrees. Applied once, to a product decision, that lesson looks like vision. Applied repeatedly, to how people are managed, it can start to look like something else.

By 1984, Jobs was running a combined division overseeing both the Macintosh and its underperforming predecessor, the Lisa. Sales were falling short of projections. Two people close to the situation tried to flag it. One was the engineer already overruled once, whose earlier memo about internal damage had gone unanswered. The other was Jobs’s own marketing manager, who proposed quietly moving him into research and development, away from day-to-day operations, before things escalated further.

Neither correction changed anything. That’s the detail worth sitting with: this wasn’t a case of information failing to arrive. It reached Jobs directly, from people inside his own working environment, through channels that were working exactly as they should. The available evidence doesn’t establish why it failed to land — only that, after two vindicated overrides, it did.

By April 1985, the correction had nowhere left to go but up. Apple’s CEO told the board he’d resign unless Jobs was removed from the Macintosh division. The board sided with him. Jobs kept the title of chairman but lost the operating role. Within months he resigned from the company entirely rather than remain in a position that was authority in name only.

It’s tempting to read that resignation as pride, or as someone unable to accept a lesser job. The record doesn’t say. What it does show is a pattern: a run of being right that was never interrupted by a cost large enough to force a reckoning, followed by a correction that arrived all at once instead of gradually.

None of this required Jobs to be difficult in some fixed, personality sense. It requ2020

ired something more ordinary: overriding consensus, winning, and not getting a second data point telling you when to stop. Most people who do this don’t get gradual feedback. They get funded, promoted, or believed — until a board, a partner, or a market delivers the correction in one move, because the smaller ones never got through.

The mechanism isn’t rare. It’s usually just invisible until it’s expensive.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

Uber China: When the Weapon Belongs to Everyone

Uber’s China strategy made perfect sense on paper. Spend aggressively, subsidize rides until rivals run out of cash, and let capital do what product usually can’t. It had worked in city after city, market after market. In China, the exact same logic bankrupted both sides at once.

Between 2014 and 2016, Uber poured as much as $1.5 billion into driver subsidies in China, part of total losses exceeding $2 billion over two years. That wasn’t recklessness. It was the company doing precisely what had made it dominant everywhere else.

The Advantage That Requires an Asymmetry

Subsidy warfare only works as a strategy if one side can outspend the other. Uber had almost always been that side. But by early 2015, Didi and Kuaidi had already merged, consolidating roughly 80–90% of China’s ride-hailing market before Uber had built comparable local scale. Uber wasn’t the incumbent bankrolling a fight against scrappy challengers. It was the challenger, and its rival had just gotten bigger.

What happened next mattered more than who arrived first. Baidu invested in Uber, including a $1.2 billion round. Apple invested $1 billion in Didi Chuxing. Two of the best-capitalized companies on earth were now backing opposite sides of the same war, at nearly matching scale. The tool Uber relied on to win — the ability to outspend — had quietly stopped being a tool it could wield alone.

The Trade That Nobody Signed

A subsidy is only a competitive advantage if the other side can’t match it. Once Didi’s backers could match Uber dollar for dollar, subsidies stopped functioning as a weapon and became something closer to rent: the mandatory cost of remaining in the match at all. Didi lost as much as RMB 12.2 billion during this period. At their peak, both companies were separately losing roughly $1 billion a year. Neither company was winning. Both were paying to avoid losing.

The distortion went further than the balance sheet. Subsidy-matching created its own incentives to game the system — drivers running dual phones to fake rides, GPS-spoofing software sold on Taobao, all designed to extract bonuses from a structure built to reward volume, not value. The war wasn’t just expensive. It was corrosive to the thing it was supposed to be building.

The Trap With No Exit Ramp

This is the shape the trap took: Uber spends to gain share. Didi’s equally capitalized backers match the spend. Neither side gains ground that holds. Both keep adding capital just to defend position, not to advance it. There was no point at which either company’s internal logic said stop — stopping unilaterally meant conceding the market immediately, and neither firm was built to concede.

Didi didn’t just defend. It went on offense elsewhere, investing $100 million in Lyft and building a loose coalition with Ola in India and Grab in Southeast Asia — raising the cost of Uber backing down anywhere, not just in China. The war had exported itself.

The Breaking Point

What finally ended it wasn’t a strategic breakthrough on either side. It came in August 2016, when Uber sold its China operations to Didi Chuxing in exchange for a 17–20% equity stake, while Didi invested $1 billion into Uber globally. The deal reportedly followed pressure from Uber’s own investors, who could see something neither operating company could act on from inside the fight: that the war was structurally unwinnable, and the only rational move was to convert a shared loss into a shared asset.

The Decoded Insight

A subsidy strategy is a weapon only when the rival can’t match it. When capital access becomes roughly symmetric, the same strategy turns into a trap that punishes both competitors equally and rewards neither — a trap neither company can exit on its own, because exiting first means losing first. Ending it takes someone standing above the fight, with a clear view of both sides, willing to call a war that neither combatant could admit was already lost.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Boundary That Made FTX Fast

For most of its short life, FTX was praised for one thing above all: speed. While rival exchanges built compliance teams and slow-moving boards, FTX moved money the way a trading desk does — instantly, informally, without asking permission. Customers loved it. So did an industry eager for a rescuer.

FTX’s new CEO, appointed the day it filed for bankruptcy, said he had never seen “such a complete failure of corporate controls.” That’s usually read as an indictment of what was missing. It’s more useful read as a description of what FTX actually was.

Why It Worked

FTX and its sister trading firm, Alameda Research, were never meant to be strangers. Alameda needed liquidity to trade; FTX needed a market maker to keep its exchange functioning. So FTX’s own code let Alameda run a negative balance — reportedly up to $65 billion — a privilege no ordinary customer had. Regular users who went negative were auto-liquidated. Alameda was exempted.

This wasn’t a hack. It was a feature, built in from FTX’s earliest days, and it worked exactly as designed: capital moved between the two companies with no wall, no ledger, no approval required. That frictionlessness is a large part of why FTX could outrun competitors weighed down by their own controls.

The Hidden Trade-Off

The advantage and the vulnerability were the same piece of infrastructure. The channel that let FTX move fast was also the only place a shortfall could hide — because it was, by design, the one place nobody was checking. There was no wall separating “the mechanism that makes us competitive” from “the mechanism that would catch misuse of customer funds.” They were never two mechanisms. They were one.

The Trap

In spring 2022, a small group of employees reviewing FTX’s code found the backdoor themselves, flagged it, and watched the concern reach Bankman-Fried’s inner circle. Nothing changed.

That detail matters more than it first appears: the organization did catch the problem, months before collapse, and still couldn’t act on it. Closing the channel wasn’t a technical fix — it meant dismantling the exact mechanism the company’s speed depended on. Meanwhile, that same channel kept absorbing more. Mid-2022 rescue loans to distressed rivals BlockFi and Voyager, funded through the undifferentiated pool, made FTX look like the industry’s last reliable player — while quietly enlarging the hole the rescues were drawn from.

The Breaking Point

The mechanism didn’t need a new failure to break FTX — it needed an audience. On November 2, 2022, a report revealed that over a third of Alameda’s $14.6 billion in assets consisted of FTX’s own token, FTT. Nothing about that fact was new to insiders. What changed was that outsiders could now see it. Within days, a rival exchange’s CEO announced he’d sell his FTT holdings, depositors rushed for the exits, and a company that had absorbed years of accumulating exposure without visible strain collapsed nine days later.

The Decoded Insight

The standard story is that FTX failed because a founder lied. The more precise story is that FTX’s growth engine and its blind spot were never separate things — the absence of a boundary was the product. That isn’t unique to crypto. Any organization whose edge is “we move faster because we skip the check” runs the same architecture: a founder-discretion culture at a scaling startup, a trading desk with a privileged internal credit line inside a larger bank, a platform where one team’s shortcuts become another team’s exposure. In each case, detection and collapse tend to arrive together, not in sequence — because the thing that would have caught the problem early was removed to make the advantage possible in the first place.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Advantage That Became a Trap

Every year for over a decade, Toys “R” Us made the same decision. Faced with a choice between funding the investments that might have kept it competitive, and servicing the debt owed to the three firms that owned it, the company chose debt. Every single time. It looked like failure. It was actually compliance.

Why It Worked

By the early 2000s, Toys “R” Us had built the kind of scale that made it irresistible — not just to shoppers, but to capital. In 2005, KKR, Bain Capital, and Vornado moved to acquire the company, putting up $1.3 billion of their own cash against $5.3 billion in debt loaded directly onto the company they were buying. The retailer’s own balance sheet would pay for its own acquisition.

The Hidden Trade-Off

The deal did more than saddle Toys “R” Us with debt. It rewired the entire logic of who got paid, and when. The company’s capital structure flipped from roughly 30% debt before the buyout to 78% debt after. Alongside that, the new owners signed themselves into a contractual Advisory Agreement worth $15 million a year, rising 5% annually — payable regardless of how the stores performed. They collected $81 million in fees tied to the original buyout debt, and more in 2009 when that debt was refinanced. None of this income depended on Toys “R” Us competing well. It depended only on Toys “R” Us continuing to exist and continuing to pay.

The Trap

This is the part that looks, in hindsight, like negligence but was really just arithmetic. Servicing $400 to $500 million in interest every year was not optional — it was the senior obligation, ahead of everything else. What was left over, after interest and fees, was what remained for the actual business: competing with retailers racing ahead in e-commerce, modernizing stores that were slowly going dark and threadbare. There was rarely much left. And because there was rarely much left, performance stayed flat or worsened — which meant even less cash the following year, which meant the same underinvestment repeated itself, again, as the reasonable choice.

None of this required bad judgment from any single executive in any single year. It required only a structure in which the party with the power to redirect capital toward survival had no financial reason to.

The Breaking Point

By 2017, more than a decade of that arithmetic had left the company with little room to absorb a bad year. When one arrived, there was no reserve of capital or credibility left to draw on, and the company filed for bankruptcy.

The Decoded Insight

It’s tempting to tell this as a story about Amazon outrunning a slow-footed retailer. But Toys “R” Us didn’t lose because no one saw the threat. It lost because the people who controlled its capital had already been paid, on schedule, whether or not the response ever worked. That’s the quieter and more durable failure mode: not an owner who makes the wrong call, but an owner whose return doesn’t depend on making any call at all. It shows up wherever a controlling party can extract value on a fixed schedule while the asset underneath them slowly runs out of the resources it needs to compete — landlords collecting rent while deferring maintenance, portfolio companies paying management fees years after the portfolio company stopped growing, any relationship where the principal gets paid first and the agent gets whatever risk is left over.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

For most of the twentieth century, Polaroid ran on a simple and extraordinarily profitable idea: sell the camera cheap, sell the film dear. Instant film carried gross margins that often exceeded 65 percent, and every camera in the world was, in effect, a standing order for more of it. This razor-and-blade logic built one of the most admired technology companies of the postwar era.

It also quietly wrote the rule by which every future decision would be judged: does this protect film revenue?

For decades nobody noticed the rule, because nothing threatened it. Then, in 1989, Polaroid put 42 percent of its research and development budget into digital imaging. By around 1990 its engineers had built a prototype camera with a 3.3-megapixel sensor, an LCD display, and memory-card storage — reportedly ahead of the first consumer digital cameras that would reach the market. By 1996, the company had working, functional digital cameras.

This is the part of the story that doesn’t fit the popular version. Polaroid was not blindsided by digital photography. It saw it coming, understood it, and built it — years before it needed to.

And then it didn’t sell it.

A digital camera creates no consumable. There is no film to reorder, no dark slide to peel away, no per-image margin. For most companies, a technology like that is simply a new product line. For Polaroid, it was a direct threat to the mechanism that funded everything else — including the very lab that had built the digital camera in the first place. Commercializing it seriously meant funding the death of the business that made commercialization possible. The company had built its own successor and then found it could not afford to let it win.

This wasn’t stupidity. Former CEO Gary DiCamillo later described the underlying belief plainly: people were betting on hard copy, on something “pick-up-able, visible, seeable, touchable.” He called it “a major hypothesis that I believed in my marrow that was wrong.” That’s the trap in miniature — not a failure to see the future, but a sincere, structurally reinforced conviction that the present would hold. The financial logic and the personal belief fed each other until neither looked like a choice.

The market did not wait for Polaroid to resolve the contradiction. Between 1995 and 2001, the business-customer share of Polaroid’s revenue fell from 50 percent to 38 percent, as digital cameras absorbed the demand instant film once owned. This wasn’t a shock; it was a slope, visible year after year, while the company kept refining the thing it already did best instead of the thing customers now wanted. By 1997 Polaroid stock traded at $60 a share. By 2001 it was frozen at 28 cents — a loss of roughly 99.5 percent of its value. That October, the company filed for Chapter 11, listing $1.8 billion in assets against $948 million in liabilities. It did not fail for lack of technology or lack of capital. It failed because the one thing it had enough of — a working alternative to its own core product — was the one thing its business model could not use.

The trap here isn’t unique to cameras. It recurs anywhere profit depends on the recurring cost of using a product, and the next technology’s entire appeal is removing that cost — a printer selling ink against a cartridge-free rival, a publisher selling subscriptions against a free and open web, any company whose margin lives inside the very friction a competitor is racing to erase. In those companies, the danger isn’t a missing lab or a slow engineer. It’s a balance sheet that has already decided which future is allowed to win.

Subscribe now

Share

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

The Advantage That Became a Trap

Uber’s board didn’t remove Travis Kalanick because he was rude. They removed him because the thing that made Uber unstoppable had stopped distinguishing between a regulator and a rape victim.

Why It Worked

Uber’s early advantage came from a simple bet: treat taxi regulation — medallion systems, city-by-city licensing, rules built for a world without smartphones — as an obstacle to route around, not a rule to obey. Software called Greyball identified regulators posing as riders and quietly blocked them, letting Uber keep operating while officials tried to shut it down. This wasn’t incidental to Uber’s growth. It was the growth strategy, and in city after city, it worked.

The Hidden Trade-Off

A company doesn’t install a reflex like that and keep it fenced to one use case. Circumvention worked so well against regulators that it became the organization’s default answer to any obstacle. Kalanick personally directed covert testing in China years before formal entry — the same instinct, now aimed at a market rather than a law. The tactic that let Uber outrun a slow-moving taxi industry was never taught to distinguish an outdated regulation from a safeguard that existed for a reason.

The Trap

That distinction’s absence turned lethal. In February 2017, a manager who had clearly propositioned an employee wasn’t meaningfully disciplined — HR called it his “first offense” and let it go. Months later, an Uber executive in Asia obtained the confidential medical records of a woman who alleged an Uber driver had raped her, in direct response to the legal threat her allegation posed. Neither case involved a regulator. Both involved a person the company was supposed to protect, met with the same avoid-the-obstacle logic once aimed at city hall.

The pattern reached into engineering, too. Before Uber’s self-driving test vehicle struck and killed a pedestrian in Tempe in 2018, the company’s cars had already been involved in 37 other incidents. Automatic emergency braking was disabled in the vehicle at the moment it hit her. Speed over caution wasn’t confined to legal strategy — it was built into how the software decided what mattered.

The Breaking Point

What broke the pattern wasn’t one scandal but several converging within months — a harassment cover-up, a stolen-trade-secrets lawsuit, an executive rifling through a rape victim’s medical file — different departments, same unscoped tactic. One insider later described the board’s prior consensus bluntly: Uber was him, and he was Uber. When investors finally forced Kalanick’s resignation, they weren’t disciplining a personality. They were trying to disable a reflex that had outgrown its original target and could no longer tell a law worth breaking from a person worth protecting.

The Decoded Insight

What changed afterward is instructive. When Uber later needed to overturn a costly labor law, it didn’t build software to dodge it — it ran a ballot campaign and won at the ballot box. When the company turned profitable in 2023, its language shifted from growth at all costs to cost discipline. The tactic hadn’t vanished. It had finally been given a boundary.

Every organization that wins by refusing to accept a rule as legitimate inherits the same unfinished work: the tactic that gets you past the gatekeepers carries no instructions for where to stop.Someone has to build that stopping point on purpose — or the instinct that broke the rule protecting incumbents will eventually break the rule protecting people.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe

For six years, Toshiba’s finance departments had a problem nobody was allowed to name. Targets kept arriving from the top that operating reality couldn’t meet. And at no point did anyone below the president’s office have a legitimate way to say so.

This wasn’t a company run by fools. In the decades after the 1973 oil shock, Toshiba scaled up centralized R&D on a simple, coherent logic: technology drives business, so direct resources toward technology and let the organization execute. It worked. A stream of world-first and Japan-first products followed — the kind of output that only happens when an organization can move fast and in lockstep, without every decision being relitigated by committee.

That coordination had a price nobody priced in: it ran in one direction only. Instructions flowed down. Nothing resistant flowed up. As long as targets were achievable, this cost nothing — speed and obedience looked identical to competence. The trouble started the moment they diverged.

In January 2009, President Nishida told the PC division to “work desperately” to improve results or face being shut down. The division complied the only way it structurally could: not by telling Nishida the target was unrealistic, but by force-selling parts to manufacturers to inflate revenue. Three years later, a demand for a ¥12 billion revenue increase arrived — this time achievable only in the final three days of a fiscal period, and only through the same kind of manipulation. The executive who issued it reportedly told investigators afterward that he’d wanted to reduce the padding. That’s not a contradiction so much as the mechanism’s clearest proof: a leader can be sincere about wanting honesty and still issue a demand that functions, structurally, exactly like an order that cannot be questioned. The system doesn’t need villains. It just needs no channel for the answer “we can’t.”

Each unmet target, once concealed successfully, became the template for the next one. By 2015, an independent committee found ¥151.8 billion in inflated profits accumulated over six years and concluded that Toshiba “had a corporate culture in which management decisions could not be challenged.” Three former presidents resigned.

The same reflex appeared where you wouldn’t expect it — outside the company’s own walls. In 2006, Toshiba had bought Westinghouse for $5.4 billion, a price analysts considered too high. Rather than revise the nuclear strategy downward, the company kept committing to an aggressive sales forecast it couldn’t walk back without admitting the original number was wrong. In 2017, the bill came due: Westinghouse filed for bankruptcy, and Toshiba booked one of the largest annual losses in Japanese corporate history.

In 2021, investigators found that Toshiba’s management had worked with Japan’s trade ministry to pressure activist shareholders ahead of a vote — one executive’s email put it bluntly: they would ask the ministry to “beat them up.” The target of the silencing had changed from an internal accountant to an external shareholder, but the reflex was identical: when someone raises an inconvenient truth, the response is to suppress the messenger, not examine the message.

By 2023, Toshiba was delisted after 74 years, taken private for $14 billion.

Every one of these episodes gets told separately — a fraud, a bad acquisition, a governance scandal. But they share one root: an organization that can only hide a wrong decision, never correct it, because nothing built into it lets a subordinate say the number is impossible. The eventual size of the reckoning isn’t set by how bad the original decision was. It’s set by how long the system can keep anyone from saying so out loud.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

Subscribe now

When Performance Becomes Reality

Enron did not begin with a desire to deceive.

It began with an obsession with performance.

That obsession helped transform a regional pipeline company into one of America’s most celebrated corporations. But eventually, something changed. Performance stopped being something Enron reported.

It became something Enron had to preserve.

The Cost of Being Right

Enron’s transformation was genuinely impressive. It moved beyond pipelines into energy trading, contracts, financial markets, and eventually EnronOnline. Revenue climbed from $13.3 billion in 1996 to $100.8 billion in 2000.

The numbers reinforced the story.

And Enron understood the power of numbers. Its 2000 shareholder letter described the company as “laser-focused on earnings per share.”

That focus was not irrational. Investors reward growth. Executives need measurable targets. A rapidly expanding company needs ambitious expectations.

The problem was what happened when extraordinary performance became the expectation.

Once a company repeatedly delivers exceptional results, merely disappointing the market can become more dangerous than making a risky decision to avoid disappointing it.

When the Number Becomes the Target

Enron increasingly developed sophisticated ways to produce, value, and structure financial results.

Some of those tools were legitimate innovations. Others became vehicles for concealing deteriorating economics.

The important point is not that Enron suddenly discovered fraud.

The problematic structures accumulated over years. SEC findings identified issues involving Chewco and JEDI as early as 1997. LJM transactions followed from 1999 onward.

The pattern was more gradual:

Performance created expectations.

Expectations created pressure.

Pressure made financial engineering more valuable.

Financial engineering temporarily preserved the appearance of performance.

That preservation then made disappointing performance even harder to accept.

The system was beginning to reinforce itself.

The Trap

This is where Enron’s failure becomes more interesting than a story about dishonest executives.

A disappointing quarter is survivable.

But if reported earnings support the company’s valuation, reputation, executive credibility, and growth narrative, then disappointing earnings threaten much more than a single quarter.

The cost of admitting weakness rises.

So the organization finds ways to postpone it.

A weak asset can be restructured. A liability can be moved. A transaction can change how an economic outcome appears. A future expectation can be recognized differently.

Each decision may solve an immediate problem.

But each solution can increase dependence on the next one.

Eventually, the organization is no longer simply trying to improve the business.

It is trying to prevent the measurement from revealing that the business is deteriorating.

When the Scoreboard Takes Over

By 2001, the gap between reported strength and underlying problems had become increasingly difficult to ignore.

Later SEC proceedings described Lay receiving information about losses, earnings shortfalls, accounting problems, and asset valuations while Enron continued publicly presenting itself as financially strong.

That contradiction reveals the deeper failure.

The numbers were supposed to measure the business.

Instead, preserving the numbers had become part of the business.

This is how a performance system can become dangerous without being obviously irrational at the beginning.

The company does not abandon its goal.

It becomes trapped by its success at achieving it.

The same pattern can appear far beyond accounting. A startup that must always grow. A salesperson who cannot miss a quota. A school judged by test scores. A government judged by economic targets.

The warning sign is not simply manipulation.

It is the moment when an organization becomes unable to tolerate what its own measurement might honestly reveal.

Because when the score becomes more important than the reality it was designed to measure, people eventually stop managing reality.

They start managing the score.

Subscribe now

Share

Share 大神仁 Hitoshi OGAMI

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription

DEC Wasn’t Too Slow to Change

Subscribe now

Digital Equipment Corporation looked, in hindsight, like the classic company that failed to adapt.

But the evidence tells a stranger story.

DEC entered personal computers. It adopted UNIX. It developed the Alpha processor. It supported Windows NT. It pursued networking and Internet businesses. It even created AltaVista, one of the early major search engines.

DEC did not ignore the future.

It kept adding it.

When Adaptation Becomes Accumulation

This was understandable.

DEC had spent decades building an enormously successful computing system around its PDP and VAX machines. Customers depended on them. Engineers understood them. Software had been built around them. Services and relationships had accumulated around the installed base.

Abandoning that system overnight would have destroyed valuable assets.

So when computing began moving in new directions, DEC had a rational response: preserve what already worked while adding what might work next.

VAX remained important.

But Alpha was developed.

UNIX was supported.

But OpenVMS remained.

Windows NT was embraced.

Intel systems were added.

Networking expanded.

The Internet became another opportunity.

Each decision made sense in isolation.

Together, they created something more dangerous.

A company increasingly responsible for maintaining several different versions of its future at once.

The Cost of Keeping Every Door Open

Optionality feels like flexibility.

Under uncertainty, it often is.

If nobody knows whether customers will choose UNIX, Windows, proprietary architectures or commodity processors, keeping multiple possibilities alive can look prudent. It reduces the risk of betting everything on the wrong future.

But options are not free.

Every new architecture requires engineering. Every operating system requires support. Every ecosystem creates relationships, products and organizational commitments. Every new direction competes for attention with the systems that already generate revenue.

Eventually, the organization faces a problem that technology alone cannot solve:

Which future deserves to win?

This is where adaptation can produce its own trap.

The organization becomes capable of responding to almost everything, but less capable of concentrating on anything.

DEC’s problem was therefore not simply that it failed to recognize technological change.

It may have been that it recognized too many changes without fully replacing enough of what came before.

When Yesterday Keeps Coming Along

This distinction matters because individual decisions can remain rational long after the overall strategy becomes difficult to coordinate.

Keep VAX because customers depend on it.

Build Alpha because proprietary performance still matters.

Support UNIX because customers want openness.

Support Windows NT because Microsoft is becoming important.

Use Intel because commodity computing is gaining ground.

Pursue the Internet because the Internet clearly matters.

None of these decisions requires stupidity.

The problem emerges from their accumulation.

Yesterday’s business remains too valuable to abandon. Tomorrow’s opportunities remain too important to ignore. So the organization carries both forward.

Eventually, adaptation itself becomes a coordination problem.

DEC was not technologically frozen in place. It was becoming strategically crowded.

The Hidden Mechanism

This is the trap of adaptation-by-addition.

When an established organization encounters disruption, the safest response is often to add the new paradigm to the old system.

That works for a while.

But repeated often enough, every new solution becomes another commitment. The organization gains capabilities while losing simplicity. It preserves options while reducing concentration.

The paradox is that an organization can become more adaptable in what it can do and less adaptable in what it can become.

That pattern extends far beyond DEC.

It appears whenever companies add AI without changing how they operate, when governments digitize bureaucracy without removing it, or when individuals accumulate new skills without changing the assumptions governing their careers.

The warning sign is not always resistance to change.

Sometimes it is the opposite.

When everything becomes part of the future, the organization may eventually lose the ability to choose one.

Subscribe now

Share

Leave a comment

Get more from 大神仁 Hitoshi OGAMI in the Substack app

Available for iOS and Android

Get the app

Refer a friend

Give a gift subscription

Get 50% off a group subscription