Think of two ways to run a coffee shop.
In the first, you brew a fixed number of cups every fifteen minutes, on schedule, whether or not anyone is standing at the counter to buy them. In the second, you brew a new cup only after someone has actually ordered one, and you only restock beans once your existing supply visibly runs low.
Henry Ford ran the first kind of shop, at the scale of an entire industry, for four decades. Taiichi Ohno built the second kind — and it took him nearly thirty years to get it right.
Ford's moving assembly line, the one already covered elsewhere on this site, pushed parts forward on a fixed schedule: the conveyor moved at a set pace, and the worker's job was to keep up with it, regardless of whether a dealership somewhere actually wanted the finished car waiting at the end of the line. It was extraordinarily fast. It was not, in Ohno's later view, actually efficient — because speed alone says nothing about whether you're building things anyone currently wants.
In 1956, Ohno traveled to the United States specifically to study Ford's methods. He came home unimpressed by the factories and oddly captivated by supermarkets — specifically, by how a shelf only got restocked after a customer had actually taken something off it. That observation became kanban: a simple card that travels backward through a factory, signaling exactly how much of something has been used and exactly how much more to make. Nothing more, and — just as importantly — nothing less.
The tidy version of this story ends there: one supermarket trip, one insight, one system. The real version is less convenient. Toyota's own internal pull-system experiments were already running years before that trip. What the visit actually did was confirm and accelerate work already in progress — which is a much better description of how the whole Toyota Production System actually came together: not a single flash of genius, but three decades of small, continuously tested revisions, most of them invisible outside the factory where they happened.
Core Philosophy
Ohno didn't invent jidoka or just-in-time production from nothing — both existed as an idea or a partial machine before he ever touched them. What he actually built, over nearly three decades of continuous, mostly small revisions rather than one grand redesign, was the connective system that made those separate ideas operate together at the scale of an entire factory. The popular version of his story compresses that slow, unglamorous process into a single clean epiphany — a supermarket trip in 1956 that suddenly revealed the whole system to him. Toyota's own internal records tell a duller, truer story: the pull-system experiments the supermarket trip supposedly inspired had already been running inside Toyota's factories for at least a couple of years beforehand. The neat origin story and the real one disagree, and the real one is the more honest advertisement for his actual method: judge a system by what it does over time, not by the story told about the moment it began.
The system took thirty years of small revisions to build. The origin story compresses it into one afternoon — which is exactly the kind of oversimplification his own method exists to catch.
How They Thought
Thinking Process
- 01
Reverse the direction information travels through the factory
Instead of a schedule pushing parts forward from raw material toward finished product, a kanban card pulls parts backward from an actual customer order — nothing gets made until something downstream has actually been used.
- 02
Make defects impossible to quietly pass along
Jidoka and the Andon cord both exist to stop a process the instant something is wrong, rather than let a flaw travel further down the line where it costs more to fix.
- 03
Treat inventory as a cost, not a cushion
Excess parts and finished goods, in Ohno's framework, weren't a safety margin — they were waste that hid other problems in the system, since a factory sitting on a large buffer never has to notice its own inefficiencies.
- 04
Turn a founder's aspiration into an operating system, not a slogan
Kiichiro Toyoda's 'build only what's needed' was a goal, not a method. Ohno's decades of work were spent building the actual mechanics — kanban, andon, standardized work — that made the goal something a factory could execute daily.
- 05
Improve continuously in small steps rather than redesign all at once
The Toyota Production System took roughly from 1948 to the mid-1970s to reach its mature form — a slow accumulation of small, tested revisions (kaizen) rather than one comprehensive plan implemented top-down.
Every one of these is about direction and speed of feedback — how fast a problem gets noticed, not how fast a part gets made.
Transferable Frameworks
Mental Models
Pull, Not Push
Let real demand trigger production backward through the system, rather than pushing output forward on a fixed schedule regardless of whether anyone needs it yet.
Stop Immediately, Fix Immediately
The cost of pausing a process the moment a defect appears is almost always lower than the cost of letting that defect travel further downstream.
Inventory Hides Problems
A large buffer of excess stock makes inefficiencies easy to ignore — shrinking the buffer forces problems to surface immediately, where they're cheaper to fix.
Turn an Aspiration Into a Mechanism
A stated goal ('build only what's needed') isn't a system. The actual mechanics that make a goal achievable daily — a card, a cord, a rule — are a separate, harder piece of work.
Small, Continuous Revisions Beat One Big Redesign
A mature operating system built from decades of small, tested improvements is more durable than one built from a single comprehensive plan imposed all at once.
Four of these describe a factory floor. All five also describe how to run a review process, a support queue, or a codebase without ever quite noticing that's what you're reading.
He Traveled to America to Study Ford. He Came Back Rejecting Ford's Central Idea.
In 1956, Ohno Toured the Country That Invented the Moving Line — and a Supermarket Impressed Him More
A Student of the Push Line
Ohno travels to the United States intending to study the Ford production system that had defined mass manufacturing since Henry Ford's Highland Park plant — the same moving line that could turn out a finished chassis every few minutes, regardless of whether a dealer downstream had actually ordered one.
Convinced the Model Itself Was Backward
He returns more impressed by how American supermarkets restocked a shelf only after a customer had actually taken something off it — an inventory system responding to real, immediate demand, not a fixed schedule. Toyota's pull-system experiments, already underway internally, accelerate from there.
Ford had already solved how to make cars fast. Ohno decided that was the wrong problem to keep solving. The question he spent the rest of his career on was how to make only as many cars as somebody actually wanted, exactly when they wanted them — a goal a faster version of Ford's line could never reach, no matter how fast it got.
The Life, Briefly
Timeline
- 1912
Born on 29 February in Dalian, in Japanese-controlled Manchuria (present-day China).
- 1932
Graduates in mechanical engineering from Nagoya Technical High School and joins Toyoda Spinning and Weaving.
- 1943
Transfers to Toyota Motor Company as a production engineer, inheriting a factory far smaller and less capitalized than Ford's American plants.
- 1948–1949
Begins experimenting with pull-based, supermarket-style inventory replenishment inside Toyota's own machine shops, years before his personal U.S. supermarket visit.
- 1949–1975
Rises steadily through Toyota's management — machine shop manager, then director (1954), managing director (1964), senior managing director (1970), and vice president (1975) — while continuously developing what becomes the Toyota Production System.
- 1956
Travels to the United States to study Ford's methods and comes back convinced the push-based model itself was the deeper problem — see Turning Point.
- 1962
Company-wide kanban implementation begins across Toyota's own plants, after years of internal piloting.
- 1978
Publishes 'Toyota Production System: Beyond Large-Scale Production' in Japan, codifying three decades of incremental development into a single account.
- 1988
The book's English translation reaches a global manufacturing and management audience for the first time, well after Toyota's methods had already begun reshaping Western industry.
- 1990
Dies on 28 May in Toyota City, Japan, at age 78.
It took him roughly thirty years to build the system his book describes in under three hundred pages.
Go Deeper
Books & Resources
Toyota Production System: Beyond Large-Scale Production — Taiichi Ohno
The primary source itself — Ohno's own account, written after three decades of practice, and the source of both quotes on this page.
The Machine That Changed the World — James P. Womack, Daniel T. Jones, and Daniel Roos
The MIT study that introduced 'lean production' to a Western audience, translating Ohno's system into terms the global auto industry could benchmark against.
Toyota Kata — Mike Rother
A closer look at the continuous, small-scale improvement habit (kaizen) underneath the visible tools like kanban and andon — the less photogenic part of Ohno's actual method.
Scholarship Notes
- Sources vary on precisely when Toyota's internal 'supermarket method' experiments began, with accounts ranging from 1948 to the early-to-mid 1950s. What's consistently corroborated is that internal piloting predated Ohno's popularly cited 1956 U.S. supermarket visit, which is best understood as an influential confirmation and refinement rather than the system's sole origin.
- Responsibility for the Toyota Production System's individual components is genuinely distributed: jidoka traces to Sakichi Toyoda's 1896 loom, the just-in-time aspiration to Kiichiro Toyoda in the 1930s, and the operational system connecting them to Ohno's own work from the late 1940s onward. This page attributes credit accordingly rather than crediting Ohno with originating every underlying idea.
Henry Ford's line moved fast in one direction: forward, on a fixed schedule, whether or not anyone downstream wanted what came off it yet. Ohno spent nearly thirty years building a system that ran the same idea backward — nothing gets made until a real order, somewhere down the line, asks for it.
The story usually told about how he got there is cleaner than what actually happened: a single supermarket visit in 1956, one flash of insight. The real record shows years of smaller, internal experiments already underway before that trip, refined slowly rather than discovered all at once.
Both the system and its origin story teach the same lesson, from two different directions: judge the thing by what it does over years, not by the moment it's convenient to say it began.