Lean is from Toyota, not Ford, and not 15th-century Venice boat builders

Anywhere but possibly inside Japan, finding local roots for Lean is useful to defuse nationalism when implementing it, but it is also risky. You start by giving a local pioneer credit  for what he actually did. Similarity of his insights with Lean then becomes enough to label him a “precursor.”  It may be a stretch, but it is a white lie, and it makes local engineers and managers so much more receptive! Further down this slippery slope, however, the local precursor becomes a “pioneer” and soon there is nothing to Lean beyond what he came up with, at which point his legacy impedes Lean  implementation more than it supports it. This is where Lean is attributed to Henry Ford.

In reality, while the founders of Toyota learned everything they could from foreign sources in early days, they and their successors are the ones who put the Toyota Production System (TPS) together and made it work, before the term “Lean Manufacturing” was coined. A Toyota alumnus told me that he never heard Toyota people claim they had invented anything; after all, they are in the car business, not the production system business. What is unique about their work is that they have integrated all the pieces — borrowed or not — into a system that outperformed the competition. As part of its 75th anniversary celebration, Toyota published the following illustration of its overall system:

From the Toyota 75th anniversary web site

From the Toyota 75th anniversary web site

They also published a detailed timeline of the development of TPS  from 1945 to 2005, highlighting the key challenges the company faced in each period, and the solutions it adopted in Just-In-Time and Jidoka. Each item has a short explanation in text, and is illustrated with cartoons, technical drawings, and photographs. It is an excellent and balanced account of the technical content of TPS, and I recommend going through it to understand how the pieces fit together.

Based on this timeline, other details contained in the 75th anniversary website, and a few other sources, I compiled the following summary, going back further in time, and emphasizing international exchanges. What I find most striking about this timeline is that the foreign inputs to TPS, primarily from the US and secondarily from Germany, were over by the mid 1950s, almost 60 years ago, and that, since the late 1970s, the flow is in the opposite direction, with the rest of world learning from Toyota.

History of Lean

TPS is still a work in progress. It has been and still is primarily an original development. The bulk of TPS has come from the minds of inventor Sakichi Toyoda, his son Kiichiro, engineers Taiichi Ohno and Shigeo Shingo, and hundreds of thousands of Toyota employees over decades. A trade secret until Toyota started training suppliers in the 1970s, TPS was revealed to the world with the publication of Taiichi Ohno’s book in 1978.

The American influence, particularly Ford’s, is readily acknowledged and played up in Toyota’s official literature. The German contribution, while not hidden, is in small print. Takt  is a central concept in TPS, and it came to Toyota from the Mitsubishi Aircraft plant in Nagoya, which had learned it from German aircraft manufacturer Junkers. After the subject of Takt came up in a LinkedIn forum a few months ago, I pulled on this linguistic thread to see what came out, and I was surprised by the magnitude of it, essentially a whole production system for aircraft, including some principles of supply chain management. It is summarized in the following blog posts:

Toyota’s study of automotive technology also included reverse engineering a 1936 DKW from Germany, and Toyota’s first postwar model, the 1947 SA, looked like a Volkswagen beetle.

Why Toyota designers chose to imitate this particular car at that particular time is another mystery, but not relevant to the key point here, which is that all of this borrowing from abroad is ancient history.

Takt time – Transfer from Germany to Japan in World War II

In Americanization and its Limits, p. 325, the sentence “In February and March 1942, visiting engineers from the German firm Junkers lectured to Japanese aircraft engineers on high-volume manufacture of fuselages and engines.” made me wonder how, in the middle of the war, engineers just “visited” from Germany. It’s not as if you could then hop on a plane in Berlin and land a few hours later in Tokyo.

According to Wikipedia, until Germany invaded Russia in June, 1941, Japanese and Germans visited each other by riding the Transsiberian railway, from Moscow to Vladivostok, which took a few weeks. After that, the only way to travel was by blockade-running submarines, and only six such trips occurred, carrying in total 96 people from Germany to Japan, and 89 in the other direction.

According to the 1946 US Air Force report on the Japanese aircraft industry, passage of materials by rail stopped after Germany attacked Russia, but passage of personnel continued. This is surprising, and it is difficult to imagine how Germans could have allowed to travel that road, but Japan was not at war with Russia until August, 1945. It had a neutrality pact with the Soviet Union, which must have allowed Japanese citizens to travel to Moscow under diplomatic cover, and then on to Germany through a neutral country like Sweden or Turkey.

Ernst Udet visited Japan, its air force and its aircraft industry in 1939. Early in 1941,  the train was still available  and General Yamashita — future “tiger of Malaya” hanged as a war criminal in 1946 — spent several months in Berlin. He brought back 250 aircraft technicians, engineers, and flight instructors. Given the timing and arithmetic, there is no way the vast majority of these 250 could have returned to Germany before the war ended.

The people who visited Mitsubishi aircraft must have been from that group, and must have been available for more than a lecture. My guess is that they stuck around to help Mitsubishi implement their Taktsystem. These people’s direct knowledge of the Junkers system is also as of 1941, before forced labor.

The JMA (Japan Management Association) currently is a large consulting firm in Japan. Shigeo Shingo worked there, and it is also where Nakajima coined the term TPM. The JMA already existed during the war, and, after the war, was instrumental in propagating techniques from aircraft manufacturing to other industries.

According to the same source, “Even in the early 1960s, the Japanese market for cars remained small and assemblers wanted to produce a variety of cars. The companies preferred to accommodate many types of cars on a single line, and consequently emphasized the need to equalize the cycle times of all production processes. This was rather similar to the situation at aircraft companies during the war…”

Takt time – Even more about origins in German aircraft manufacturing

Earlier this week, I ran into John Paxton’s 2008 paper called Myth vs. Reality: The Question of Mass Production in WW-II, in which he makes a convincing case that production methods were far more advanced in the US aircraft industry than in Germany or Japan. It is really not in doubt. The point in trying to understand the Junkers Taktsystem is simply as one of the sources of TPS. World War II German and Japanese engineers could design advanced planes, like the first jetfighter, the Messerschmitt 262 that you can see in the Smithsonian today, or the Mitsubishi Zero. But, in production, they could not come anywhere near the one-bomber-an-hour performance of Ford’s Willow Run plant.

Yesterday, I was able to go to the main library at Stanford University, where they have about one foot on one shelf in the basement with books on the German aircraft industry in World War II. including in particular Lutz Budrass’s work on the subject and Holger Lorenz‘s Kennzeichen Junkers. Budrass’s book is a forbidding 1,000 pages of small print with a few grainy pictures, long on armament policies and politics, but short on technology:

Lorenz’s book is much more accessible and contains many high-quality photographs, which contradict Paxton when he says:

“Photographs from the era show this  difference. Classic photos from Vickers and DeHavilland (British) and Junkers and Heinkel (German) production facilities show  isolated aircraft in ‘final assembly’, in  stationary jigs, being assembled by ‘work  gangs’, embodying the ‘craft production’  process. In contrast, photographs from  Grumman, North American, Republic, and others show rolling final assembly processes, with aircraft moving from station  to station, much like Model T assembly  twenty years earlier.”

What Paxton writes is not consistent with what little I have seen about  the Junkers Taktsystem, and it is not consistent with the photographic evidence either, which shows that, at Junkers, the final assembly methods in the 1930s were eerily similar to those today for airliners, as you can see in the following side-by-side comparison:

Junkers clearly had a rolling assembly line, albeit one that, unlike the Boeing 737 line, was not continuously moving, Otherwise, the similarities extend even to the kits of parts that are rolled over to each plane. Of course, this is Junkers, not the whole of the German aircraft industry at the time, but Junkers is the one we are interested in, because of their Taktsystem and their transfer of this method to Mitsubishi aircraft in Nagoya, through which it reached Toyota. Other companies used a variety of methods. As we can see on the following picture Heinkel 111 bombers appear to have been assembled on fixed stations in 1939, but Messerschmitt fighters on assembly lines in 1943:

These and more pictures of German aircraft manufacturing before and during World War II are available from the Bundesarchiv Picture Database.

Upstream in fuselage assembly, the comparison looks as follows:

In 1934, for Ju-52 fuselages, Junkers used a nose-to-tail assembly line; in 1940, for the Ju-88, a side-by-side line. Today, Boeing 737 fuselage assembly, Spirit Aero in Wichita, KS, appears to be using parallel fixed stations. Fuselage assembly in this context, however, is limited to fastening together sections that have been assembled separately, with automated riveting.

Paxton’s article contains other assertions that are also difficult to accept. He claims, for example, that the abundance of cars in the US spread mechanical skills throughout the population and that these skills made it easier for large numbers of workers to learn how to build airplanes. He quotes the following statistics for the number of people per vehicle in different countries in 1926:

  • Australia:  30
  • China: 31,871
  • Japan: 1,789
  • Britain:  49
  • France:  54
  • Germany: 194
  • Italy: 353
  • United States: 6

The US was the only country in which almost every family had one car, and American cars of that era were designed to be maintained by their owners. They came with a kit containing the necessary tools and instructions. The US aircraft industry during World War II, however, is known to have employed women in large numbers, as in the following photo of  women installing fixtures and assemblies to a tail fuselage section of a B-17F Bomber (Library of Congress).

If do-it-yourself car maintenance pre-trained World War II aircraft workers in mechanics, then car maintenance must have been done by women. Single women in isolated farms certainly had no choice but to maintain  their cars and tractors, but, in the culture of the 1920’s and 30’s, it is difficult to imagine that women who had a man at hand didn’t delegate changing spark plugs to him rather than learn it themselves. Paxton’s article also asserts that German aircraft production required skilled craftsmen, but most of it during the war was done by forced laborers from occupied countries who were not trained mechanics or machinists.

On the other hand, the article fails to mention two obvious reasons for the superior performance of aircraft manufacturing in the US:

  1. Among all the belligerents, the US was the only one with aircraft factories that were out of the enemy’s range. Germany’s ABC program, on the other hand, had defense against air attacks as a central design consideration. Whatever you do to spread out the facilities and protect the supply chain from bombs does not help you in productivity or quality. The Junkers factory in Dessau was bombed in 1944.
  2. The US aircraft industry in World War II had a highly motivated work force. Not only were these manufacturing jobs the best these women had ever had, but they knew they were producing equipment for the men in their lives who were fighting for a cause they believed in. By contrast, in Germany, Rosie the Riveter’s counterparts were Polish or French workers building bombers against their will for the worst thugs in history, and they would have been happy to see these planes crash on take-off. In addition, most of them didn’t speak German.

The picture that emerges from the documents I have seen so far is that, in the late 1930s, Junkers had organized production in what is now called pulse lines. Final assembly was divided into operations of balanced durations, so that the planes didn’t move during operations but moved forward in unison at fixed time intervals, with upstream processes and the supply chain organized to support this mode of operation. And this is the Taktsystem that was taught to Mitsubishi Aircraft by Junkers engineers in 1942.