Was NUMMI a Success?
By now you know that Toyota made its much-anticipated decision to close NUMMI. Many of my friends are saddened by the turn of events. While I am also sad, I’m also okay with the decision. All good things come to an end, and if NUMMI was to ever cease operations, now is a good time.
My work with NUMMI was the greatest experience of my professional career. My learning curve was so steep I couldn't see to the next step, let alone the top of the stairs. I remember thinking, literally, "If someone came along and made me rich with a big, fat check, I would still do exactly what I’m doing now." And what I was doing then was working incredible hours and, if not loving every minute of it (it had its less than pleasant aspects as well), appreciating every moment.
So, a question I've heard a lot recently: was NUMMI a success? Rather than simply reply with a simple yes or no, let me share my own experience there.
Toyota hired me in late 1983 to work on the Toyota side of its new venture with GM. I was assigned to a newly formed group at the company's Toyota City headquarters to develop and deliver training programs to support its looming overseas expansion.
All this was just happening. NUMMI didn't even have a name yet. The agreement with the UAW was yet to be signed. (One of the first meetings I attended was an explanation of the just-signed Letter of Intent. I was relieved they didn't ask me to translate - the content of the presentation by the US attorney was far too technical for the level of my Japanese language at the time. The final agreement wasn't ratified until summer of 1985.) There weren't yet any employees of NUMMI, not even management. NUMMI wasn’t successful, it wasn't famous - it was just a dream.
GM had several very clear business objectives. They had an idle plant (the Fremont plant had made several products over the years, Chevy Camaro, Olds Ciera, GMC trucks. But, no matter the product, quality was consistently GM's worst. Worst of the worst.) Not to mention an idle workforce, the fact of which was never helpful to the company's overall relations with the UAW.
More to the point of a venture with Toyota, GM didn't know how to make a small car profitably and partnering would give them a chance to see how Toyota does it. And it would give them a chance to see Toyota's production system. The Toyota Production System wasn’t famous yet, but there were people in GM who knew a little about it and wanted to know more.
What did Toyota want?
The specific objectives for Toyota were less tangible. But it was crystal clear that Toyota needed to manufacture cars in North America. Political pressure made it no longer tenable to simply ship cars around the world from the comfortable confines of Toyota City. And Toyota was trailing Honda and Nissan, each of who had already established operations in Ohio and Tennessee.
So Toyota needed to produce in the U.S. But why California, the most expensive place in the U.S. to manufacture (NUMMI paid about $13 per hour in the beginning, a little on the high side for the industry at that time)? Why with GM, still by far the largest car company in the world? And why with a UAW workforce??
Using an existing facility would obviously save time and money. And above all it would provide a cheap way to learn quickly.
That's what Toyota wanted most of all: to learn. GM had a clear set of business objectives above all. Toyota, on the other hand, had a different way of approaching this venture.
What exactly did they want to learn? Above all, Toyota faced two big unknowns when it came to operating outside Japan (even outside Toyota City): People and Suppliers.
Toyota had a lot of confidence in its system. Between 1950 and 1980, the company had evolved a way of working that was revolutionary. They were confident in their ability to physically put together all the mechanical pieces of producing an automobile. But the people and supplier parts were scary.
The Toyota Production System starts and ends with people building quality into the process. All the way back to Toyoda group founder Sakichi Toyoda's early 20th century automatic loom. Sakichi’s loom was brilliant in the way it combined automation with people. The automation was made to work FOR the people - not the other way around - in pursuit of better quality. Sakichi's system engaged workers minds as well as their hands in identifying and responding to problems, developing effective countermeasures to the root causes of problems on the spot. Quality wasn’t inspected in - it was built in.
And, secondly, Toyota's just-in-time system of achieving end-to-end material velocity depended on close working relationships with suppliers. NUMMI's suppliers would be expected to deliver with absolute reliability.
People and suppliers, those were the difficult questions and challenges -- how could they be answered quickly? Countermeasure: GM and the UAW as partners.
It is no secret that before approaching GM Toyota held discussions with Ford about the possibility of entering into joint production. Toyota had long held great admiration for Ford, admiration from the Toyoda family for the Ford family, and admiration by Toyota the company for Ford the company as well. But, Arab oil embargo concerns combined with a general lack of interest on the Ford side led the talks to quickly fizzle.
So, instead of Ford, GM was the easy second choice. GM is no big surprise, but what about the UAW?
It would be going too far to say that Toyota in the beginning actually wanted the UAW as a partner in the venture. But, once it became clear that the UAW was going to be in the picture, Toyota embraced them fully as a partner: "If we can make NUMMI a success with a UAW workforce in California, we can be successful anywhere."
On the supplier side, it is well-known that Toyota grew up with its own semi-captive set of keiretsu suppliers. Working with new suppliers was always a serious matter for Toyota, demanding great care. As NUMMI's first general manager of production control (and later president and chairman of Kanto Auto Body) Susumu Uchikawa said, "Without our suppliers, we can’t do anything." He was just acknowledging the truth. OEMs rely on suppliers for most of the components that go into the final product. With automobiles, everything, every component, is engineered for each vehicle. Totally integrated engineering. Close partnership with suppliers just makes good business sense. You can't produce with good quality and profitability if your suppliers are weak and going out of business. You don't get rich being a Toyota supplier, but neither do you go out of business!
GM as partner (Ford would have worked as well) could introduce their suppliers and show how they work together. Learning those lessons turned out to be a long, difficult process. But, NUMMI and GM's help was the beginning of Toyota's long journey in this critical learning process.
Learning the people side of things, especially how to work effectively with American front line production workers, proved a much faster course. While it wasn't easy, it was remarkably successful; Toyota, NUMMI, the UAW, and the entire workforce, achieved Toyota City levels of performance through attainment of an extraordinary degree of mutual trust. In terms of tangible performance, NUMMI didn't just improve; it went from GM's worst plant to its very best. That improvement was achieved in just one year, and with the same workforce.
The NUMMI Learning Ledger
I left Toyota's employ in 1994. Let’s look at the Ledger of Learning at that time, ten years into the joint venture:
Learn about North American suppliers - check
Learn how to work with North American people - check
In terms of specific, tangible results, GM had indeed benefited and arguably much more so than Toyota. The original product from NUMMI all went to GM in the form of the Chevrolet Nova. So, as planned, NUMMI enabled GM to add a nice small car to its line-up. A new model even formed the flagship product for a new brand (a sub-brand of Chevy, really), the Geo Prism in 1988.
Toyota on the other hand, initially received no product at all from NUMMI. A new hatchback version of the Corolla was added a couple of years after start-up, but quickly bombed in the marketplace (the Corolla FX - no penalty points if you don't remember it). Eventually, Corolla sedans and the small pick-up in 1991.
So that's the way the ledger looked in the mid-90s, half-way through the life of the JV. Note that the JV wasn't even supposed to last this long to begin with. Back in the beginning, Chrysler led a lawsuit to prevent NUMMI from even getting off the ground, claiming monopoly concerns. Here were the largest auto companies in the US and Japan teaming up on everybody else. The judgment handed down limited the life of the JV to 12 years, so NUMMI should have closed up shop in 1996. But, as the time approached, an appeal to the Court resulted in a ruling that allowed the JV to operate with no legal time limit.
As it turns out, legal time limit or no - the end is in sight.
If that was the Learning Ledger half-way through, what about now, 25 years in as the venture comes to an end?
Toyota got the basic learning it wanted very early on. Toyota's ongoing involvement as GM's JV partner at NUMMI has been a matter of loyalty as much as anything. Under ordinary circumstances, Toyota would never close an operation it had invested in. Toyota has a track record of proving time and again the lengths to which it will go to preserve jobs well past the apparent business need for them.
From the beginning, Toyota's objectives at NUMMI were defined by learning rather than the specific business tangible objective that typically define a joint venture. And if there's one thing Toyota knows how to do it is how to learn, especially where it's important down at the operational levels of the company - a characteristic that is the embodiment of the learning organization. Toyota's biggest strength is that it learned how to learn, and it was that approach to learning that defined its approach to NUMMI from day one.
For GM, on the other hand, it was only about half-way through the life of the JV that the deeper learning started to pay off. Jack Smith was on the negotiating team that created the agreement with Toyota. As chairman in the mid-90s, Smith is the one who finally put senior level shoulder into making something of the learning of the by then substantial number of mid-level managers who had gone to NUMMI to learn. With the edict to "run common, run lean," Smith authorized a team to execute the establishment of GMS - Global Manufacturing System - and to build a model lean factory where all the learnings of the previous ten or so years could be put together. They decided it would be easier to experiment far from the mothership in Detroit (and away from the UAW) so chose Eisenach in Germany. From there, plants in Brazil, China, finally back in the good old USA (even in the middle of Michigan and with the UAW!) incorporated the same principles, design, and ways of working.
GM's global initiative was successful because of the deep learning that had occurred among the ranks of people with NUMMI experience. In the 1980s, GM NUMMI grads recast the TPS they had learned at NUMMI into something they called "Synchronous Production" and began offering training to significant numbers of GM people. Meanwhile, back at NUMMI, GM turned its small liaison office - called TLO, or Technical Liaison Office - into a training operation that organized short and longer-term visits to NUMMI into true development opportunities for the GM folks who went through there.
It all finally paid off, by many objective standards and according to numerous third-party observers, GM’s new plants are world class, in quality and cost.
And, starting about five years ago, GM even began applying what it had learned of lean practice in the plants to work in the office. GM’s "Enterprise GMS" initiative saved a billion dollars in the first year of applying lean thinking to office processes.
But, by 20 years or so into the venture, GM seemed to have decided that it had really had enough of NUMMI. They recreated the training provided by the NUMMI TLO by offering the same experience in Michigan, eventually shutting down the training operation at the liaison office at NUMMI.
And, by the end, when other GM operations had improved so much, NUMMI no longer provided the dramatic impact on visitors it did back in the 80s, when visitors would leave slack-jawed at what they had seen. In fact, by the end, not only was NUMMI no longer GM's quality leader, NUMMI was actually a drag on GM's overall quality scores. That means: the quality of products manufactured at NUMMI was worse than GM’s average! Ouch. (Note: that is my understanding based on discussions with various people - I can’t verify it, but it sounds credible.)
So, why is NUMMI closing?
Clearly, the dollars and cents don’t add up for either Toyota or the new GM. Neither needs the capacity right now, and who’s to say if they will need it in the future.
But, also - more importantly - the learning is done. Fini. Caput.
Or so it seems.
But, I would argue that there is still a LOT more to learn. About technology transfer, the dissemination of learning, the MANAGEMENT system that underpins and enables the more famous Production System, the importance and attainment of mutual trust between labor and management, about how to sustain a powerful operating system over decades and decades.
NUMMI was a great story in its own right. A story of people coming together and doing something great at a time and place in history. And NUMMI was important to both GM and Toyota. I think it was important for the UAW, too.
But, a less considered fact is that NUMMI was hugely important to American manufacturing. NUMMI proved that the best, supposedly "Japanese," production methods in the world could work on American soil with American labor. An early motto at NUMMI was "Best of Both Worlds." I truly believe NUMMI in its heyday embodied that motto in principle and in practice. And, if lean production and lean thinking and the lean enterprise are the way forward for American organizations in all industries, NUMMI was the most important lens for the world outside Toyota to see it up close.
John Shook, Senior Advisor
lean enterprise Institute, Inc.
Time To Make Time
When the people in a lean system don't value time, everyone is cheated, says John Shook, in this fascinating reflection on the role that time plays in a close observation of work.
The Remarkable Chief Engineer
How can a system in which "we are all connected and no one is in charge" support purposeful and productive work? Toyota's famed Chief Engineer system has much to offer in this regard. John Shook explores how the leadership styles of, and ways of working by, the CE might provide something of a roadmap for all of us.
How Standardized Work Integrates People With Process
In this three part series on SW, John Shook argues that "the Toyota Way is a socio-technical system on steroids. A test for all our lean systems is the question of how well we integrate people with process (the social with the technical). Nowhere does that come together more than in the form of standardized work and kaizen."