NUMMI joint venture in California was a classroom for change

From Toyota, GM learned the importance of consistency and communication

In the early 1980s, General Motors quietly negotiated an unprecedented deal: GM, the world's largest automaker, would build cars jointly with fast-rising challenger Toyota.

Both parties had something to gain from the talks, which created New United Motor Manufacturing Inc., known as NUMMI. Toyota wanted to learn to build cars in the United States.

In a sense, GM wanted the same thing. Lulled by years of dominance in the United States, GM had allowed lean Japanese competitors to surpass it in manufacturing efficiency. GM needed to learn how to build cars in the United States, too.

"If you looked at the domestic manufacturers, GM plants were competitive," says Don Hackworth, retired GM senior vice president. "But when you put them up against the best in the world, they fell short."

What GM leaders didn't initially grasp was how pulling on the string of manufacturing inefficiency would unravel much of their corporate structure — and culture. As executives studied GM's competitive shortcomings, including high costs and mediocre quality, they realized that the GM empire had to change radically.

Eventually, GM responded to the NUMMI challenge. Enlightened executives, eager to reform GM manufacturing, marketing and other functions, eliminated many corporate fiefdoms and adopted efficient practices worldwide. But GM executives now agree: It took too long.

The path to NUMMI

GM's path to NUMMI was preceded by what Hackworth calls "a slow and insidious recognition that we were falling short."

It was a sensitive topic for the proud — some say arrogant — company. Ron Harbour recalls the reception his father, Jim, got in 1980 when he initiated the Harbour Report, an annual study of industry productivity.

"He made this presentation of the findings, and basically GM told him, 'You don't know what you're talking about. We're the biggest in the world, and therefore we're the best,' " says Harbour, a partner in Oliver Wyman's global automotive practice. "They summarily dismissed those numbers."

But, he adds, "What we found out later was that GM did take it seriously and did many of their own studies. If anything, we were probably understating the problem."

Fritz Henderson, GM COO, says manufacturing wasn't uniformly bad. But efficiency varied widely from plant to plant.

Too many variations

"We did have a whole list of operations that were quite productive," Henderson says. "In the past, oftentimes the performance of a plant, the work force of the plant, was a function of the plant manager. You had so-and-so's manufacturing system or so-and-so's philosophy."

Without a common system, Henderson adds, "Everybody was relearning lessons all the time."

Hackworth sees such variations as the core problem of GM's old system.

For example, in 1994, GM had 75 stamping-line configurations. The company concluded at the time it needed only six. Plus, it had some stamping machines working just 10 hours per week. Far better are standard but versatile machines working continuously.

The disciplines involved in designing a vehicle — engineering, design, manufacturing, powertrain — also were poorly coordinated. The design studio would style a car and "throw it over the fence" to manufacturing, which had to figure out if the car could be built.

Battling divisions

But GM had an even more fundamental problem.

The car divisions, fabled empires in GM's decentralized business structure, made it virtually impossible for GM to compete. Hackworth, 71, who ruled one of those empires as general manager of Buick from 1984 to 1986, says the last thing the divisions wanted to do was share best practices. Instead, they were intensely competitive.

"They were all fighting for scarce resources, and they all had their own little skunkworks, and I put myself right in that category, you know," Hackworth recalls. "I was developing a two-seater that no one knew about. That's the way you did things."

Manufacturing techniques that might provide an edge over other GM divisions were hoarded, he says. During GM's era of domestic dominance, its leaders believed that such rivalry sharpened the divisions' performance. But stiff competition from the Japanese forced GM to rethink that.

"How can you say that Oldsmobile was the enemy when Toyota was knocking on our door?" Hackworth says.

Surprises at NUMMI

NUMMI started production in 1984, reopening a shuttered GM plant in Fremont, Calif. GM's NUMMI personnel got a close look at Toyota's vaunted production system. What they saw surprised them, recalls retired executive Bob Hendry, 64. Hendry served as NUMMI general manager of general affairs, later becoming chairman of Adam Opel.

"Some of the learning was that there wasn't a magic bullet that the Toyota production system had," Hendry says. "Whatever this thing was didn't have any magic in it. When we brought manufacturing executives to consult with them, I often heard: 'Yeah, we do that.' "

But, he says, Toyota plant workers followed best practices consistently — eliminating the variation that plagued GM.

Mark Hogan, 57, former GM executive who also served as NUMMI general manager of general affairs, says Toyota used more manual labor than GM expected. GM saw NUMMI workers doing considerable manual welding, for instance. At a time when GM leadership saw robotics as the key to productivity, this was a revelation.

Toyota's efficiency came from making plant workers more productive. Hogan says that Toyota's standardized procedures helped workers improve on the job: "Those were very foreign concepts, but when you step back and look at it, I mean, how logical, so logical."

Hogan recalls that while GM plant supervisors ran 30 to 35 people, NUMMI teams were five or six people. Also, he says, NUMMI workers "were taught to be essentially industrial engineers" who did multiple tasks. UAW workers at GM plants worked within rigid job classifications.

Ultimately, GM executives realized that matching Toyota would require companywide change. A unified, coordinated system would have to replace the fiefdoms that ran GM product development and manufacturing. As Hendry puts it, "People look at the Toyota production system as a manufacturing system. What we learned was that it was a total company system."

Enter Jack Smith

Recognizing the lessons from NUMMI was one thing; forcing change throughout GM was quite another.

Harbour says efforts to spread the lessons of NUMMI initially "sputtered and didn't go anywhere." For years, the manufacturing system stifled enthusiastic NUMMI veterans.

"They bring him back and throw him in a GM plant and he's one of 3,000 people," Harbour says. "All of a sudden he looks like the zealot among the 2,999 other people, and they all look at him like he's some kind of freak."

That pattern changed after the 1992 boardroom coup that put Jack Smith atop GM. Smith, now 70, who helped negotiate the NUMMI deal, says the lessons from NUMMI "were a wake-up call in Detroit."

Harbour credits Smith with laying down the law. GM began assigning clusters of NUMMI alumni to plants. Change began to accelerate.

"By the late '90s, they had traction and were making significant improvement," Harbour says.

He says that the UAW bought in when it learned that making plants more productive sometimes meant that GM would move outsourced work back in-house. Harbour recalls being in a GM plant when stamping dies were brought back into the plant.

"There were union people standing around applauding," Harbour says. "That was a big, emotional event."

Running common

Under Smith and his successor, Rick Wagoner, GM began a push to "run common." GM created single global organizations for engineering, design, manufacturing and powertrain.

GM's divisions market and sell cars, but no longer have their own secret projects. Vehicles are developed under vehicle line executives who run multidisciplinary teams. The teams include manufacturing engineers — no more hard-to-build designs tossed "over the fence."

Plants around the world have common layouts, processes and equipment. Gary Cowger, 61, GM group vice president for global manufacturing and labor relations, says that makes it easy to transfer plant managers or other key personnel.

"When they walk in, they see the same processes," Cowger says. "They see the same systems, and they understand exactly what needs to be done to continue to have that plant, wherever it is, operate efficiently and with high quality."

Although Cowger warns that chasing efficiency is "a never-ending process," GM has seen a payoff from its wrenching effort. The company's productivity numbers, as measured by the Harbour Report, are among the best in North America. In the 2007 report, GM had three of the top 10 assembly plants in productivity.

Jack Smith says GM's progress traces back to lessons learned at NUMMI.

"Now, admittedly it took us a long time to get that learning, to implement those concepts modified for General Motors in our systems," Smith says. "But eventually it got done. I just wish it had happened a lot faster."

You can reach Dave Guilford at dguilford@crain.com. -- Follow Dave on