When Efficiency Removed the Human 効率が削ったもの

Over twelve million American small and mid-sized businesses now face what economists call the "Silver Tsunami" — the largest generational transfer of business ownership in modern history. Within ten years, roughly seventy percent of these companies will change hands. Yet, by most accounts, fewer than a third have any documented plan for what comes next.

This is not, properly speaking, a retirement problem. It is a continuity crisis. Behind the statistics lies a pattern that has destroyed every organization, in every era, that allowed itself to become predictable.

Tech giants lay off tens of thousands while their stock prices rise. Startups that raised hundreds of millions quietly close. Companies that seemed invincible — fortress brands, locked-in customers, cutting-edge technology — are quietly counting the years they have left. They are all walking the same path. The names change. The pattern does not.

The Five Stages

Every organization rises and falls along the same five stages.

Stage 1. A charismatic founder emerges. The vision is singular. The decisions are fast. Apple under Jobs. Amazon under Bezos. Walmart under Walton. The energy is undeniable, the rules few, the possibilities limitless.

Stage 2. Systems are built. The founder's intuition becomes process, documentation, culture. The organization scales beyond the founder. This is called maturity. It is also, in another sense, the beginning of the end.

Stage 3. The founder dies, retires, or is replaced. The source of original judgment is gone. What remains is the system that the founder built — perfectly tuned for the world that no longer exists.

Stage 4. Succession struggle and internal division. Factions form. Politics intensifies. The best people start to leave. The company spends more energy fighting itself than fighting its competitors. Most of the great corporate stories of organizational decay live here — Intel, IBM, GE, Sears, the post-Jobs Apple in some readings.

Stage 5. Talent exodus and collapse. The organization is acquired, broken up, or quietly shrinks into a husk of what it was. The customers lose faith. The talent has already left.

Most struggling organizations alive today, by my count, are in Stage 4.

This was not the failure of stupidity. These organizations had access to the best MBA graduates, the most advanced technologies, the deepest pools of capital in human history. They had every advantage. They still followed the script.

The Wrong Cure

Today, faced with this pattern, leaders turn to artificial intelligence. The reasoning is straightforward: if past judgment failed, perhaps better judgment — colder, faster, more comprehensive — can be machined. Strategy meetings now routinely begin with "let's see what the model says." Hiring is increasingly mediated by algorithm. Pricing, allocation, product roadmaps — all are slowly being transferred to systems that promise the optimal answer.

This approach, taken alone, will not break the pattern. It will accelerate it.

Artificial intelligence learns from the past. It analyzes vast historical archives and offers solutions that have, on average, worked before. Those solutions are not new futures. They are highly refined repackagings of the past.

If artificial intelligence had advised Blockbuster in 2000, it would have recommended optimizing the store-location algorithm and the late-fee schedule, because those were the variables the historical data spoke to. It would not have predicted streaming, because streaming barely existed in the training set.

If artificial intelligence had advised Nokia in 2005, it would have recommended improving hardware manufacturing and reinforcing carrier relationships, because those were the existing axes of competition. It would not have seen the iPhone.

If artificial intelligence had advised Borders in 2008, it would have recommended optimizing physical retail. It would not have predicted the e-book revolution.

The general principle is uncomfortable. Artificial intelligence is brilliant at fighting yesterday's war. The decision landscape has already moved.

The Other Cost

There is a second, deeper problem with optimization as a strategy.

Efficiency, taken to its limit, removes the very randomness that produces innovation.

Look at the current wave of corporate restructuring. Companies are cutting costs based on data-driven analysis. The model says, in effect: these departments are unprofitable, these roles do not generate immediate return. What the model cannot see is which of those "unproductive" people would have created the next breakthrough, which "inefficient" team would have built the next disruptive product, which conversation in which hallway would have started the next decade of growth.

Gmail came out of Google's twenty-percent time — unstructured, unjustified hours that no rational allocation system would approve. The Post-it Note came out of a failed glue experiment at 3M. Viagra came out of unexpected side effects in a heart medication trial. Every meaningful breakthrough I can find in the history of business carries this signature: it was not on the plan.

The next breakthrough never comes from optimization. It comes from deviation.

And deviation, by definition, looks wasteful in advance.

What Lean Removed

The twentieth century achieved what may be the most refined operational philosophy in human history. Taiichi Ohno at Toyota in the 1950s defined a category that James Womack and Daniel Jones later named, in their 1990 study, Lean. It went on to reshape global manufacturing, then software, then nearly every modern enterprise.

Lean has a vocabulary. At its center is a Japanese word: Muda. Waste. The seven Muda. The eight Muda. The endless varieties of waste that good systems should eliminate.

Lean was right, within its own frame. Inventories shrank. Defects vanished. Throughput multiplied. Capital efficiency improved by orders of magnitude.

And yet, in March of 2011, when the Tōhoku earthquake and tsunami struck Japan, Toyota's Lean system — perfected over fifty years — collapsed in days. Single-source suppliers had been chosen for cost. Buffer stocks had been eliminated as waste. Slack had been engineered out of the system. When one node failed, the entire global production line stopped.

Then, in 2020, the same thing happened on a global scale. Lean supply chains everywhere — in masks, in chips, in food, in cars — collapsed under shock. The same word kept appearing in the after-action reports: just-in-time became just-too-late.

The lesson is structural, not historical. Perfect efficiency, taken to its limit, produces perfect fragility.

What Lean called Muda — what it eliminated — was not merely waste. Some of it was something else.

Reading Muda Differently

The book you are now reading proposes a different reading of the same word.

Yes, some Muda is waste. Defective inventory, redundant motion, excess transportation — Lean is right to eliminate them. No serious operator disagrees.

But there is a second category of Muda that Lean, by its very structure, cannot see. Buffer capacity. Redundant suppliers. Unstructured time. Conversations without an agenda. Relationships without immediate transactional value. The long-arc decisions that produce no measurable return for years. The judgments that resist quantification.

None of these can be priced. None of them survive a quarterly review. Most of them are invisible to artificial intelligence by construction. And yet, on inspection, every long-surviving organization — sacred sites that have stood for thirteen hundred years, family enterprises that have lasted eight hundred, companies that have outlived every competitor across centuries — preserves precisely these.

What they preserve is what Lean removes.

This book gives that preserved thing a name. I call it Muda — using the same word Lean uses, but pointing to the other half of what the word can mean.

The Muda of Lean is what efficiency removes. The Muda I describe in this book is what survival requires.

The proposition of the book, stated as cleanly as I can:

Muda is the human. Judgment, relationship, the long arc, the unmeasured. What twentieth-century management called waste. What twenty-first-century AI cannot see. What every organization that has survived more than a few generations has insisted on keeping.

From Lean to Muda

The arc of this book is suggested by its subtitle: From Lean to Muda.

This is not an attack on Lean. Lean built the twentieth century's productive base. I have no intention of unbuilding it. But the twentieth century ended. The next era will not be won by competing on the same axis that Lean optimized. That axis is now in the hands of artificial intelligence, and on that axis, no human enterprise can win against algorithmic optimization at the limit.

The competitive edge of the next era lies elsewhere. It lies in the capacity to keep, and to read, what efficiency systematically removes. It lies in the human.

This is the move I call From Lean to Muda.

A Note on the Origin

The methodology I lay out in this book is not, in its essential structure, new. Eight centuries ago in Japan, a Buddhist monk named Ano Zenjo (1153–1203) had already articulated something very close to it. Yoritomo and Masako, the founders of the first samurai government in Japan, were his disciples. Recent historical research, drawn from primary sources, has established that the visible governance of the Kamakura Shogunate stood, in considerable part, on the invisible network that Zenjo built. The visible depended on the invisible. The institutional form depended on what could not be institutionalized.

That work, and that man, were systematically erased from the official record. The textbook history we inherited tells a different story, in which Zenjo plays a marginal role. The case for restoring the original picture is laid out, in primary-source detail, in the academic essays at tokiwatanabe.org. I will sketch it briefly in Chapter 5 of this book.

I want to be exact about my own position here. I am not the heir to this method. I am its translator. Ano Zenjo himself did not use the word "Muda." What he called Ura — the hidden, the invisible, the back of the cloth — I have translated, eight centuries later, into the language of twenty-first-century management as Muda. The naming is my responsibility. The thing itself has always been there.

The Promise of This Book

By the end of this book, you will have the following.

A framework for identifying which of the Five Stages your organization is currently in — and what the typical failure pattern of that stage looks like.

A vocabulary for the four dimensions of Muda — judgment, relationship, long-arc, unmeasured — and how to recognize each in your own organization.

A reasoned argument, from first principles, for what artificial intelligence is structurally unable to see, and where the human therefore retains an irreducible advantage.

A set of case studies — Ise Grand Shrine, Kongō Gumi, Stora Enso, Patagonia, Costco, In-N-Out, the Nintendo Wii, Toyota's hybrid-and-hydrogen bet, the Kyoto restaurants that refused their Michelin stars — of organizations that survived, in some cases for over a millennium, by keeping what efficiency told them to remove.

And a set of practices, at the level of the individual career, the organization, and the larger institutional landscape, for keeping Muda alive in your own context.

The book is short for what it attempts. It does not need to be long. The argument is, in the end, simple. The twentieth century perfected the visible. The twenty-first century is delivering machines that will perfect the visible more completely than any human could. The competitive edge of the era ahead is on the other side of the same word.

From Lean to Muda.

Let me show you what I mean.