(FORTUNE Magazine) – SATURN WAS GOING TO BE SOMETHING WONDERFUL. WHEN GENERAL Motors first came up with the idea of a company within a company, Saturn would make superb little cars to beat the Japanese at their own game. The cars would be distinctive--they'd have plastic, dent-free bodies. Dealerships would be warm, friendly zones free of haggling. Customers would find all this irresistible. And so would the rest of GM. Saturn wouldn't merely blossom as a division with protected status, free from the labor strife, stifling bureaucracy, and all the other dysfunctions of the mother corporation. No, it would also infect the rest of the company with its enlightened and effective management techniques. This was the little car that would revitalize General Motors. As the famous ad campaign proclaimed back in the early 1990s, Saturn would be "a different kind of company."

It really was a noble idea that had (and has) tremendous appeal. The idea was, as vice chairman Bob Lutz says today, "Why can't we have it both ways? Let's have wonderful dealers and consumers who are enthusiastic about the product." And some of the original concept even came true. After its first car went on sale in 1990, Saturn did go on to set new standards for satisfied buyers and to form remarkably strong bonds with its customers. "The Saturn experience," which started with the sale at the dealer and ran through the life of the car, became the talk of the industry.

But other than that, things haven't gone as planned. In fact, Saturn has been something of a fiasco. "I view Saturn as a failure," says George Magliano, director of automotive industry research for Global Insight. "It hasn't done what they expected it to do. Saturn has had two or three incarnations, and none of them has worked."

In 14 years of operation, Saturn has built a gigantic new factory, introduced three all-new models, conducted massive ad campaigns, and cost its parent company billions of dollars. Meanwhile, the notion of a tiny independent company making low-margin small cars proved totally unworkable, and competition from Japan, and later Korea, has proved--big shocker here--much tougher than anyone expected.

In all, GM has probably plowed a grand total of $15 billion into Saturn--and the division has not earned a dime. Even for a company of GM's size, $15 billion is a lot, and it's money the company can ill afford to waste. Its auto operations in North America and Europe are both in the red, and it has run up enormous unfunded retiree pension and health-care obligations. Investing that money somewhere besides Saturn could have filled some big holes. If GM had just kept the money in the bank, it might have been able to boost its credit rating above its current soggy level. (For its part, GM says the $15 billion figure is "significantly high" and that the cash spent on Saturn wouldn't have made a big dent in its other costs).

Here's the surprise: After studying the facts, GM has decided ... to up the ante. It's committing another $3 billion on a mission to destroy Saturn in order to save it. GM has taken away the special status and made it into a plain-old division of the mother ship, with the same centralized engineering and marketing as all the other brands. Unlike the original models that were developed by Saturn engineers--not to mention designed by Saturn designers and built in dedicated Saturn plants--the new vehicles will come from GM's development system and will be built in GM factories. Over the next two years the marque will get a version of a Chevrolet minivan, a sedan based on the same components as Chevrolets and Opels, and a tweaked version of a Pontiac convertible sports car. Not only will Saturn no longer be a different kind of company, it won't even be a different kind of car.

All of which still leaves a nagging question: Why bother? Isn't this throwing good money after bad? After all, Rick Wagoner, GM's chairman and CEO, likes to say that the simplest way for GM to boost its profits is to stop selling cars that lose money. Saturn, with its tide of red ink, is a huge money-loser at GM. Why doesn't GM just save itself the trouble and put Saturn out of its misery?

The reason is that Saturn has life in it yet. The Saturn name is arguably stronger than some others at GM--namely, Buick and Pontiac. Saturn regularly ranks alongside Lexus and Infiniti in terms of customer satisfaction. And Saturn owners are extremely loyal to their dealers, who requite the love with a 30-day return policy, easy service appointments, and other public displays of affection. If GM were to pull the plug, it would have to buy out that network of beloved dealers, which would also cost hundreds of millions of dollars. At least with the $3 billion in new money GM is spending, there's a possibility of profit down the road.

Saturn is also a wedge into a world of car buyers GM can't otherwise reach. According to the company's own research, 70% of Saturn buyers don't consider buying another GM car. As vice president Jill Lajdziak reminds skeptics, Saturn is a "channel to a larger ocean."

If the plan succeeds, Saturn will retain for GM those coveted affluent buyers who'd otherwise buy imports, and it will do so with a broader product line that costs less to develop. And if company projections prove true, this de-Saturnized Saturn has a chance to break even before 2010.

TURNING SATURN INTO ANOTHER GM division is exactly the opposite of what its founder had in mind.A generation ago, Roger Smith, GM's chairman and CEO from 1981 to 1990, set out to make over the world's largest automaker. Smith recognized that GM had become bureaucratic, insular, and dysfunctional, and dedicated his nine-year tenure as CEO to shaking it up. Early on, he tried to figure out why GM's small cars couldn't compete with Toyota's and Honda's. He decided that to be successful, GM needed to slough off all its old ways of doing business and start fresh.

Smith was also a sucker for big ideas, especially those involving technology. And so in 1983 he announced the creation of a special small-car project named after the rocket that carried Americans to the moon. Early on, it aimed to sell 500,000 cars a year.

As the biggest new thing out of Detroit since Ford introduced the Edsel in 1957, Saturn created a sensation. Every byte of new information that dribbled out was seized upon by a waiting world. Finding a site for Saturn's new plant turned into the biggest nationwide search since David O. Selznick sought an actress to play Scarlett O'Hara. When a rural location outside Nashville in Spring Hill, Tenn., was chosen, thousands of people showed up to apply for jobs.

Saturn started production on July 30, 1990, selling a pair of four-cylinder small cars, a sedan, and a two-door coupe. Sales were slow at first. The dent-proof plastic panels were hard to make and created unsightly gaps, and the Saturn look already seemed dated by the time it was introduced. The sedan resembled an Oldsmobile Cutlass Supreme.

But the "Different kind of company" campaign, created by the Hal Riney agency, made a splash. The ads cleverly focused more on the rural assembly plant and the down-home people who worked in it than on the vehicles themselves. Most were unabashedly corny. One featured a 10-year-old boy and his dog, who were uprooted from their Midwest home when the boy's father moved to work for Saturn, and the family began a new life in friendly Spring Hill. The intent of the ads, which never mentioned GM, was to attract import buyers who otherwise wouldn't shop for GM cars.

Problem was, most Saturn buyers never traded up. While Saturn achieved its goal of attracting buyers who weren't typically interested in GM cars, it didn't change their opinions of the company. When owners sold their first Saturn, they typically bought a second one, or they shopped elsewhere. In fact, though nobody recognized it at the time, Saturn's popularity peaked around 1994.

Saturn's troubles began when Smith's handpicked successor, Robert Stempel, was driven from office in a coup by GM's board of directors in October 1992. That left Saturn with no friends at the top of the company. Other GM divisions, particularly Chevy, were jealous of Saturn's special treatment and successfully battled it for scarce capital. Saturn executives felt as if they had been disinherited. "The genius of Roger's plan was he put us off to the side," says Don Hudler, a Saturn veteran who served as president from 1995 to 1998 and now owns six Saturn dealerships. "That cost us a lot because a lot of people in GM hated us for that. It is like being a very young child born into a wealthy family. We were getting some of their money."

While Saturn was being starved of resources, its executives inexplicably decided that the time was right for a foray into the most intransigent market with the pickiest customers on earth: Japan. So in 1997 it began building right-hand-drive Saturns and shipping them to new standalone dealerships in Tokyo and other cities. Saturn aimed to sell 3,000 to 4,000 cars its first year, but Japanese consumers, who had access to the best small cars in the world, bought only 1,400 during the first 16 months. Saturn executives talked bravely about staying in Japan for the long haul after steep losses, but they abandoned the effort in 2000.

Meanwhile, GM sacrificed another division. In December 2000 it killed Oldsmobile, which had once sold more than a million cars per year. Should GM have let Saturn die instead? "That's a moot point and is best suited to a stimulating but fruitless intellectual debate," says Bob Lutz. He argues that Olds overlapped with other divisions while Saturn gets a different kind of buyer, "and that's of huge strategic importance."

Problems piled up. Press reports said Saturn was losing $850 million in 2000 and would lose the same amount the following year. A small SUV called Vue, introduced in 2001, was initially well received, but was later found to have suspension problems; another car called Ion that came out a year later got only a lukewarm response. Increasingly, pieces of the original Saturn were stripped away in the interest of efficiency, and the once-independent company was slowly integrated into GM. Something needed to change.

JILL LAJDZIAK, WHO HAD BEEN WITH Saturn since the beginning, had long watched the decline with dismay. Having risen to become its top sales and marketing executive in 1999, Lajdziak, now 47, believed that Saturn's original concept had reached a dead end. Making plastic-bodied small cars in dedicated plants would never produce sufficient economies of scale. For Saturn to survive, Lajdziak (pronounced LAY-jack) believed it needed to boost sales volume, cut engineering and manufacturing costs by borrowing more GM resources, and develop additional products in higher-profit segments by adapting existing GM designs.

Lajdziak got a chance to put her ideas into action when Lutz arrived at GM. At a dinner after the January 2002 Detroit auto show, Lutz asked Lajdziak and other division heads for ideas on how to improve their performance. Lajdziak sent back a multipage letter outlining new product and marketing ideas. She had nothing to lose because GM had totally ignored Saturn in its future product planning. Saturn was expected to limp along indefinitely with its lineup of three vehicles.

Lajdziak won Lutz over, and he assigned Ed Welburn, GM's head of advanced design, to create some new styling themes for the brand. The diplomatic Welburn, who had served a two-year stint at Saturn in the 1990s, would become GM's chief designer in 2004. His first effort, "doesn't reach far enough," Lajdziak declared, and Lutz backed her up. "Jill made it clear that it wasn't the right vehicle for the brand," Welburn recalls. Lajdziak thought the mockups were too conservative. She wanted something daring--something that would push this perennial design laggard to the vanguard. So Welburn solicited proposals from seven of GM's 11 global design studios, and relied especially heavily on Opel in Russelsheim, Germany, where he had worked in the late 1990s. Some Opel designers moved to Detroit to work on Saturn.

GM's board of directors were briefed on the new model program as it progressed, and last summer they approved the expenditure of some $3 billion on Saturn through 2007.In fact,the way GM brass sees it, there's nothing wrong with Saturn that a few new models won't cure. "We're investing in Saturn's future because the inherent health of the brand is quite good," says Lutz: "It just needs a bigger, more exciting product portfolio."

The main victim of the new Saturn regime is engineering distinctiveness. The 2005 Saturn Relay "sport van," just reaching dealers, will be marketed under different names by three other divisions: Chevy, Pontiac, and Buick. Beneath its new sheet metal is an old GM minivan that has been on the market since 1997. Saturn's roadster, named Sky, which arrives in early 2006, looks fast standing still and will be available with a turbocharger to make it go even faster. Still, it is built from the same architecture as Pontiac's upcoming Solstice. Some analysts wonder whether selling hot cars to sensible Saturn buyers isn't akin to marketing tournament snowboards to grandmothers.

Another thing buyers will notice about the new models is that the plastic body panels are gone. Saturn used to promote the panels heavily because it had little else to sell, featuring them in commercials that showed them fending off dents from trash cans and bicycles. Some had argued that they were integral to Saturn's brand identity, but there was no room for them in GM's complex global product-development system.

Lutz says tapping into GM's global resources "is exactly what will deliver the variety of product Saturn needs." But with Saturn sharing components with other GM divisions, there is a danger that it may cannibalize other models' sales. Saturn executives say they expect to boost its volume from the 215,000 units forecast for 2004 to 400,000 in 2007. Since GM's market share has been declining for 40 years, a bump up in Saturn volume could mean a downturn for other GM brands.

Not that those other brands necessarily deserve to live, and indeed there's plenty of speculation around Detroit that GM, even after the demise of Oldsmobile, hasn't finished thinning its herd. The company won't comment on plans to mercy kill other brands, nor has it given any indication that it plans to do so. But it could certainly make sense to eliminate Buick, and possibly Pontiac as well. That would leave GM with trucks (GMC), cars for the masses (Chevy), luxury vehicles (Cadillac), and a line that aims directly at people whose tastes turn to Hondas and Toyotas and other imports. Not a bad lineup.

So Saturn hasn't been a complete disaster; it's more of a $15 billion learning experience. "Would they ever do a Saturn again in the current situation?" asks David Cole, head of the Center for Automotive Research in Ann Arbor. "No, I don't think so." Saturn is dead. Long live Saturn.