(FORTUNE Magazine) – The incident is forever seared in Jack Smith's memory. When the former General Motors CEO was still a fast-rising executive in the early 1980s, he visited Japan to study Toyota's stamping and assembly operations--something nobody at GM, amazingly, had done before. After collecting the data, Smith was astonished to discover that GM needed more than twice as many people as Toyota to build the same number of cars. But when he made his report to GM's executive committee, the council of elders that ran the company, he was met with utter disbelief. They dismissed his findings. Recalls Smith: "Never in my life have I been so quickly and unceremoniously blown out of the water."

GM was in denial. In fact, GM was so far in denial that nobody had bothered to look for a simple explanation: GM was organized in a completely different way from Toyota--as well as from Ford, Volkswagen, and every other automaker. That was a big reason GM performed so poorly. In the early 20th century GM had been assembled from independent automakers--Chevrolet, Oakland (later Pontiac), Oldsmobile, Buick, and Cadillac--all of which did things differently and, 60 years later, were still competing ferociously with one another. Only a gigantic bureaucracy kept this unruly confederation from destroying itself. Toyota started out in 1947 building cars with a single brand in one factory. As it got bigger, it just grew out from the center. Headquarters kept making all the decisions, and everybody did things the same way.

So while Toyota was unified and centralized, GM was diffuse and decentralized. GM might market eight midsized sedans--all with different names and different parts--in half a dozen markets around the world; Toyota usually marketed only one or two. Even if nobody else did, Smith realized that GM's decentralized structure was fundamentally flawed. So when the board of directors installed him as CEO in 1992, he set out to change the organization. "I had the opportunity to really structure the business in the way I thought it should run," Smith recalls in a telephone interview with FORTUNE from his winter home in Naples, Fla. "Frankly, at the time we weren't like any other auto company in the world."

Smith and his successor, Rick Wagoner, spent the next decade making over GM in the image of its formidable Japanese competitor. If it were a TV show, you'd call it Toyota Eye for the GM Guys. Smith and Wagoner pushed, pulled, pleaded, and sometimes shoved until they reshaped their unwieldy organization into a single company that was able, in Smith's words, to "run common and lean." Today the engineer who specifies the design of a wheel cover at GM's operation in Shanghai uses the same computer language as his colleagues in Germany or Michigan, sources the piece through the same global purchasing organization, and sees it installed at a factory that follows the same GM system in use on five continents.

The final pieces of this consolidation were still coming together early this year, as 4,000 engineers scattered around southeastern Michigan were moving their offices into GM's Technical Center in Warren, Mich., joining some 14,000 colleagues. For the first time in its 96-year history, GM is running as one company. Wagoner, who took over as CEO in 2000, likes what he sees. "Operating-wise, we are doing pretty well," he says.

Wagoner should know, as he has planted himself in the driver's seat and keeps a firm grip on the controls. Soon after becoming president and chief operating officer in 1998, he created a management committee, the Automotive Strategy Board, that keeps him in the face of every one of his top operating executives. Every month he gathers all 14 of them for a day-and-a-half-long meeting. The sessions are usually held on the 38th floor of one of five tall tubes at the old Renaissance Center in downtown Detroit that now serves as GM's headquarters. Dates for the meetings are set a year in advance. Attendance is mandatory. The heads of regional operations in Europe, Latin America, and Asia attend in person or by videoconference. In weeks when no meeting is held, one-hour conference calls with the same cast of characters provide updates.

Each meeting follows the same format. Before noon on the previous Thursday, presentations and supporting material are posted on a secure website. By the weekend participants are expected to have reviewed the material and made comments. The first day's session runs from 8 A.M. to 6 P.M. It is mostly business--budgets, product programs, and sales performance--but the discussions are spirited. The second day focuses on process (improving productivity, efficiency, and quality) and personnel. "You can't undersell the importance of all that melding together," says Wagoner. "It is critical to the way we run today." Adds Smith: "All the players are there. The old General Motors didn't look like that."

At this point it would be nice to report that post-makeover GM is delighting customers, dazzling investors, and blowing the doors off the competition. It isn't. In the all-important U.S. market, GM's performance remains a mixed bag. It enjoys spectacular results with full-sized pickup trucks and derivatives like the Chevy Suburban and the Cadillac Escalade that sell in huge volumes (1.7 million annually) and produce outsized variable profits--as much as $15,000 per vehicle for some versions of the Escalade. But GM loses money on small cars marketed by Saturn and Chevy, and it has shortchanged Pontiac and Buick, leaving them with outdated models. Overall, GM managed to squeeze a profit margin of only 1% out of North America last year. It had to lavish incentives on buyers to keep its market share stable so that it can fund its enormous legacy costs: $87 billion for pensions, more than $60 billion for future health-care obligations. But those incentives also eroded profits. As GM steps up its product development efforts--it is launching some two dozen new models this year--it expects to both increase its market share and reduce its reliance on incentives.

In Europe and South America, GM is losing money. Its biggest cash generators are the finance company, which has been riding the home-refinancing boom, and China, where GM moved quickly to establish joint ventures. GM now earns nearly half as much in Asia ($577 million last year) as it does in North America ($1.2 billion). All of that is reflected in the stock, which, at about $45 a share, trades just 50% higher than it did in 1992, when Smith took over.

If you strip out the legacy costs--the price companies pay for their longevity--the picture looks brighter. Analyst Himanshu Patel at J.P. Morgan estimates that GM's underlying operating margin in North America is "an impressive 5.5%," making it the most profitable mass-market carmaker outside Japan. But it still lags behind German luxury-car makers BMW and Mercedes-Benz, as well as Honda, Toyota, and Nissan.Wagoner blames some of GM's problems since 1992 on driver error, but others he attributes to potholes along the way. "We didn't do everything right over the last 12 years. About three times during that period you think you've got it, and then something else comes up. We made huge progress in '92, '93, '94, but then we started to pay the price because we had, out of necessity, underspent [on new products] during that tight period and didn't do that well on the revenue side. Then, in the last three or four years, we've run the business very well, but the pricing has been tough. So my lesson--these aren't complaints but an observation--is, Don't ever think you've got it licked, because you probably don't. This is not a one-step game. This is a multiple[-year] thing, and it's hard, and you learn as you go along. The key thing is to remember what you've learned so you don't have to relearn it."

Garel Rhys, who studies the auto industry from Wales, where he is the director of the Centre for Automotive Industry Research at Cardiff Business School, thinks Wagoner and Smith deserve plenty of credit. "When you look at a company that became No. 1 in the world in 1931 and is still there--that is an achievement, notwithstanding the difficulties they went through from the 1970s until a few years ago. Basically, the company is still the one to be shot at. It will be a very interesting battle for supremacy between GM, the old titleholder, and the new contender, Toyota. It is the only car company that has the potential to overtake them."

Fifty years ago gm didn't have to worry about competition. What the railroads and oil were to the second half of the 19th century, the automobile was to the first half of the 20th, and no auto company was mightier than GM. Having passed Ford in car sales in 1931 to become the world leader, it cruised through World War II and into the 1950s like a majestic ocean liner--a leviathan whose wake touched nearly every American. It wasn't a question of what was good for General Motors being good for the country: GM was the country. By itself, GM generated revenues equivalent to 3% of the gross national product, and that didn't include the contribution of its dealers or suppliers. (Today GM's revenues equal about 1.6% of GNP.)

In 1954, GM accounted for more than half the cars sold in the U.S. Chevrolet alone churned out 1.7 million vehicles, nearly a quarter of what rolled out of showrooms that year. On the first FORTUNE 500 list in 1955, GM was No. 1, with registered sales of $9.8 billion--42% higher than runner-up Standard Oil. Ford, still a private company, didn't make the list; Chrysler appeared in the No. 6 position, with sales about one-quarter of GM's total. GM was also No. 1 in net income, with $806 million, producing a not-too-shabby profit margin of 8.2%. In January 1955 former spark plug executive Harlow Curtice, who had become GM's president and CEO, was named Time's Man of the Year.

GM's success in the 1950s still rode on the famous aspirational ladder created by Alfred P. Sloan, in which customers would first buy a Chevy, then move up to Pontiac, Oldsmobile, Buick, and finally Cadillac. But making the organizations behind the brands work together was a huge problem Sloan never tried to solve. Each had begun life as an independent company, and Sloan had decided that each needed a degree of operational freedom to be innovative and responsive to the market. To keep them from colliding, he set up a decision-making structure of policy committees and staffs. He called it "decentralized operations and responsibilities with coordinated control."

Sloan's system served GM well as long as the car market kept growing and the competition remained weak. But by the 1980s it had outlived its time. Japanese automakers were attracting customers with better-made, less expensive cars, while GM coasted on the residual strength of its famous brands and product lines like full-sized trucks, where imports didn't compete. Inefficiencies caused by the use of unique parts and processes throughout the company were huge. Some units, such as Fisher Body, sold parts to the rest of GM for cost-plus prices that weren't remotely competitive. Meanwhile, all those policy committees and staffs slowed decision-making. By one estimate, it took five months to resolve even a tiny matter like changing a dealership franchise. The issue would have to make its way from a car division to the corporate product policy group, then to the executive committee, and finally to the board of directors.

GM had tried to change its ways in 1984, and the effort failed so spectacularly that it traumatized the company for a generation. An astonishingly candid account of what went wrong is laid out in General Motors: Priorities and Focus, a book published in 2000 for internal consumption. The mere idea that GM would encourage an airing of past failures so infuriated one retired GM chairman that author William Pelfrey no longer gives interviews, and copies aren't readily available.

The 1984 reorganization pushed through by then-CEO Roger Smith was a classic example of the law of unintended consequences. It was meant to streamline GM's processes by merging the engineering and manufacturing activities of the five passenger-car divisions and Fisher Body into two centralized organizations. One was called Chevrolet-Pontiac-Canada, or C-P-C for short, and the other Buick-Oldsmobile-Cadillac, or B-O-C. The scheme was concocted by McKinsey consultants, which should have been a warning flag, because instead of simplifying operations, as Pelfrey wrote, "sadly, it had the opposite effect." The two groups had overlapping responsibilities but shared few processes, and the informal communication networks that actually made the company run were destroyed. Complexity mushroomed and quality suffered, because the reorganization also wiped out Fisher Body, which had imposed some Toyota-like common standards on manufacturing.

The reorganization signaled the onset of a doleful period for GM. "The '80s were probably the toughest decade in GM's history from an automotive perspective," says Wagoner. "The financial results and market share results weren't that bad. But it seems like we lost some of our relative competitive positioning during the decade. We hired a lot of people, and we've been carrying those people since. We've been 12 to 14 years digging out from that."

Jack Smith, who was running GM's Canada operations at the time, avoided the chaos by moving to Europe. There he boosted profits by knitting together GM's operations across the continent and forming a single purchasing organization. So when Smith took over the parent company in 1992, he knew immediately what he wanted to do. "We were run by divisions and countries," he recalls. "Our major competitors were run with one engineering center, one selling effort. Boy, we needed to get that kind of organization." Chastened by the 1984 disaster, Smith decided to move slowly and deliberately to get there. "It was a question of how much you are going to rock the boat," he explains.

His first target, purchasing, was relatively small, and he was able to act fast. No fewer than 27 purchasing organizations operated in North America, buying everything from steel and rubber to paper clips and toilet paper. GM was paying a premium for complexity and losing economies that come from volume buys. Within a month Smith centralized purchasing into one global operation.

Next came the giant manufacturing and engineering groups, a unification process that would stretch out for a decade. Gary Cowger, president of North American operations, compared the task of modernizing the factories to tuning up an engine while the car is moving. GM couldn't economically retool a plant for common processes until a new model changeover was coming, and then it had to execute both jobs smoothly. "We had to continue to develop cars and continue to launch cars," says Cowger. "You couldn't just stop and say, 'Okay, marketplace, we're going to take two years off and get this thing realigned.' At the same time you almost continually had the thing in a reorganization mode because you were continually adding groups together."

Ditto for engineering. Putting its organizations together was like assembling a jigsaw puzzle, with each move requiring between 18 months and two years. Before two engineering groups could be joined, a physical location had to be selected and common processes like computer systems installed. In 1993, Cadillac in Detroit and the luxury-car group in Flint were merged. Two years later that group was combined with the midsized car group in Warren. Detroit-based Saturn and the small-car group in Lansing were next. And soon after, all the car operations were consolidated into one organization. Finally, in 2000 the car and truck operations were unified into a North American engineering unit.

If any sparks flew from pushing thousands of engineers together, they weren't visible from the outside. GM wasn't as lucky with its marketing divisions. Crunching Chevy, Pontiac, Olds, Buick, Cadillac, and GMC Truck into one group produced fireworks that could be seen for miles. Marketing people tended to wear their brand affiliations on their sleeves and feel closer ties to Chevy or Olds than they did to GM. Besides, each unit had an important external constituency--its dealers, who reacted loudly and publicly to any change that threatened to undermine their status.

Like most mergers, this one made theoretical sense: Put all the vehicle-marketing divisions under a single staff to eliminate duplication, internal competition, and brand overlap, replacing the six people who visit a dealer with a single person representing all the brands. It sounded sensible, except the dealers didn't like the change. The owner of a Pontiac store didn't want to be called by the same rep who was selling cars to the Buick store down the street. He was afraid some of his competitive secrets might go down the street too. "It took them a while to get used to that," says John Middlebrook, vice president of marketing and advertising.

It didn't help that GM had also found other ways to irritate dealers. It installed a computerized system that required them to order cars up to 90 days in advance and made changes difficult. Sales tumbled one month when the system collapsed. The computers got fixed, but GM was also trying to take over local dealer cooperative advertising to coordinate the ad effort and get volume discounts. And it decided to go into competition with its own dealers by announcing plans to buy as many as 700 outlets and run them itself. "In hindsight, those were both mistakes," says Middlebrook.

Managing the new marketing organization was that rarest of GM executives--someone who hadn't risen through the ranks and had no previous auto-industry experience. Ron Zarrella was working at optical company Bausch & Lomb when GM recruited him to bring in some fresh marketing ideas. Zarrella had lots of them and had a habit of saying them out loud, which didn't win him many dealer friends. Wagoner takes part of the blame. "Was some of that [Zarrella's] fault?" he asks. "Sure. But did we give him the right mentoring, did somebody step in and say, 'Be careful about going the wrong way with the dealers?' Maybe guys like me or Jack didn't do enough."

When the heat got too intense, CEO Smith was called in to put down the dealer rebellion, standing up at a meeting in January 2000 to make a public apology. "Rick stood behind Ron," says Smith. "But that's not to say the dealers weren't happy when he left [in 2002, to go back to Bausch & Lomb]. They didn't like Ron. The car business is so unique. You can come in as an outsider and look at some things and say, 'That's really a dumb way to run.' But the unfortunate thing is, that is the way it runs. There are some things you just have to live with."

Smith stepped down as CEO in June 2000 and chairman in May 2003, handing over both jobs to Wagoner. Young for a GM chief executive at 51, Wagoner has some of Smith's qualities--low-key but charismatic leadership, little visible ego, a focus on results. But he's less of a traditionalist, and he shook up Detroit when he brought in two veteran executives from rival car companies in 2001 to serve as his highest-ranking subordinates: former Ford CFO John Devine and former Chrysler vice chairman Bob Lutz. By force of personality and reputation, Lutz has made GM's product development efforts more efficient, basing decisions on experience and common sense and reducing the company's reliance on focus groups and market research.

GM still needs to move faster and smarter. One lesson it hasn't learned from Toyota is how to consistently develop successful new models that share common components. Toyota can build five or six vehicles based on the Camry sedan and make hits of them all. GM has been trying to do something similar for a decade, developing cars from a single set of key components that can be sold not only in North America but also overseas by foreign affiliates like Saab, Opel, and Suzuki. Too often it has wound up with ungainly cars because design, engineering, and manufacturing weren't using the same playbook. The latest casualty: In late February, GM decided to end production of the midsized Saturn L300, a year ahead of schedule. Designed with the same underpinnings as a model from Opel, GM's European affiliate, the L300 had been selling poorly since it was introduced in 1999. California-based analyst George Peterson says the interior was too cramped and the design too spartan for American tastes.

The L300 follows earlier flops like the Opel-based Cadillac Catera sedan and the Chevy-Pontiac-Oldsmobile-Opel minivan. To design a van that would fit on European roads, GM made it narrower than competitors'. But the minivan never caught on in Europe, and it struggled in the U.S. too. Says Middlebrook: "If you compromise a vehicle for a specific market--if you compromise the van in terms of width--then you wind up with one that doesn't work that well in either market. Now that we have more flexibility in our plants, we know how to have more wheelbase or more width in a common platform."

Wagoner believes gm has learned from its mistakes. "What went wrong," he says, "has to do with everything from organizational mentality and organization structure to how different markets really are and the capability to execute products that can be sold in more than one market. My sense is we've matured in all these areas. Probably most important, the competitive dynamics are clear to all of us." He has high hopes for cars like the 2004 Pontiac GTO, which is based on a rear-drive vehicle developed in Australia by GM's Holden affiliate but is powered by a traditional American V-8.

Making a success of common systems is essential to GM's survival, but the details can be tricky. Planners have to drive economies of scale at the same time they encourage individuality to attract different sets of customers. While that means mastering myriad details, it can produce big payoffs. "You get savings in both product and process," says engineer Mark Hogan, who runs advanced product development. "If you can double your volume, it is very attractive to suppliers. If we can buy a million steering gears instead of 500,000, it can mean a 20% cost reduction."

It hasn't escaped Wagoner's attention that while GM is becoming more like Toyota, Toyota has changed too: It is becoming more like GM. GM is integrating its local engineering and design operations in places like Germany and Brazil; Toyota is opening design and engineering centers in North America and Europe to more closely tailor its products to local tastes. Wagoner can only hope that the savings GM gets will offset any additional success on Toyota's part. "We are moving in the opposite direction," he says. "[But] we are moving to this so-called sweet spot, the happy medium." He calls it a "race to the middle."

While Wagoner doesn't pine for the good old days, he smiles when talking about the GM of 50 years ago. "I definitely [would] like having 50% of the [U.S.] market. If you look back to GM in the '50s, GM in the '60s, you get 10% net income margins for a couple of years. In the '70s, when I joined the company, you had maybe 42% market share and 5% to 6% net margin. It would be less than honest to say it hasn't been a difficult 20 years." (In 2003, GM's share of the U.S. market was 28%, and its corporate net margin was 2%.)

The number Wagoner watches as closely as market share is market cap. At the end of 1954, GM's was $8.5 billion--$58.8 billion in today's dollars. That's more than double today's $26 billion. Says Wagoner: "In that sense, we have been pretty consistently punished, not inappropriately, and ultimately that is going to be [a measure] of whether we can get this business back to where it should be. Can we get that valuation of the company up to those multiples? That is not an easy assignment."

Now that both GM and Toyota are each lean and in sync, success will go to the company that can leverage its strengths most effectively. To Wagoner that means pitting GM's design innovation and its knowledge of local markets against Toyota's smooth product development and robust balance sheet. Wagoner sometimes likes to remind listeners that GM founder William Crapo Durant lost his fortune in the 1920s speculating in the stock market and at the end of his life ran a bowling alley in Flint, subsisting on handouts from Sloan and others. Wagoner jokes that Durant's fate "has haunted GM chairmen ever since." Well, probably not. But his comment is yet another sign that GM is no longer in denial.

FEEDBACK ataylor@fortunemail.com