(FORTUNE Magazine) – You don't become head of General Electric by being a scaredy-cat, but one thing frightens Jack Welch, even after 18 years in the job: the sheer size of the company. "Don't talk to me about how big this place is," he says, nearly shouting. "I hate it." Size means sloth, systems, and smugness rather than speed, simplicity, and self-confidence. Over and over, Welch urges people to "tear this place apart."

At the same time, of course, he has grown GE so much it's, well, scary. GE is the ninth-biggest and second most profitable company in the world. Since Welch took over in 1981, GE's sales have risen 3.7 times (from $27.2 billion to $100.5 billion), and profits have grown 5.7 times (from $1.6 billion to $9.2 billion). Size is a burden if you try to lug it around, but it's an advantage if you can make it work for you. "Jack always says, 'You can't manage a company the size of GE. You leverage it,'" says Claudio Santiago. Spanish-born Santiago, 42, works in Florence, Italy, as CEO of one of GE's bigger divisions, a $2.1-billion-in-sales compressor and turbine manufacturer, Nuovo Pignone.

Welcome to global GE. Businesses like Nuovo Pignone symbolize one of the biggest stories of Welch's storied career: how America's most admired company, and a very American one, has become a truly global corporation. It is the story of how GE leveraged its businesses to buy and build, in just over a decade, a European business nearly as large as all GE was when Welch became CEO, and now intends to do the same in Asia, only faster. It is the story of how GE leverages its intellectual capital--an incomparable management tool kit that has turned a bunch of state-run sinecures and niche businesses into a powerhouse whose net income, now a fifth of the company's, is compounding at 34% a year. And it's the story of the lever itself, the reason this Gargantua can be as agile as Nijinksy: a leadership-development system that won't let GE stop improving.

As Welch prepares to retire next year, much attention has focused on the race to succeed him (see Staying Smart). He'll be a tough act to follow: If the creek don't rise, Welch will have created more shareholder wealth than any CEO in history. But his real legacy is the tools and leaders he has helped to forge, and nothing displays it better than GE's globalization.

Until a decade ago, GE treated the world as an export market. The company had some European operations, notably its aircraft-engine manufacturing joint venture with Snecma of France, formed in 1974, and a greenfield plastics operation begun in 1970 when Welch ran that business. But in 1987, 44% of GE's overseas sales were exports from the U.S. That year Welch gained a European base for GE Medical Systems by swapping his consumer electronics business for Thomson-CGR, the medical equipment division of France's Thomson, then state owned. Two years later, as the Iron Curtain rang down, GE landed another government-owned business, Hungary's lighting manufacturer, Tungsram. Since 1990, GE has paid nearly $30 billion for 133 European acquisitions; some 90,000 people are on GE's payroll there, about as many as Euro-giants Axa and BP Amoco employ worldwide. Europe accounted for $24.4 billion of 1998 revenue, of which only $2.7 billion, or 11%, came from exports from America. Inflation-adjusted, that's less than at the start of the decade.

GE's first buys were anything but aggressive. In lighting, "we were on the defensive," John Opie, now vice chairman, recalled in 1994. Dutch giant Philips bought Westinghouse's U.S. business in 1983; Germany's Siemens was making goo-goo eyes at GTE Sylvania (buying it in 1993). Tungsram was a mess--good lamps hidden under a bushel of communist mismanagement--but had 7% of the European market. If it fell into other hands, Thomas Edison's lighting business would be trapped, with strong rivals on its home turf and no way to counterattack on theirs.

Those early buys were stubbornly unprofitable, partly because they were defensive, partly because GE was still unsure how to export management know-how. Says Mike Zafirovski, a Macedonian who in July was named CEO of GE Lighting worldwide: "From '92 to '94 almost no one could convince Jack to buy in Europe. He was losing money at Tungsram, Capital, everywhere--Europe was a dirty word." GE turned to the attack only when signs of a turnaround came and a European recession made assets cheap. Between 1994 and 1995, European revenues jumped from $9.1 billion to $14.1 billion as GE made big buys in everything from reinsurance to power controls.

When you do a lot of deals--GE made 108 acquisitions worldwide last year--you get good at it. One secret is an "integration model" developed chiefly at deal-hungry GE Capital Services. It rests on an obvious fact, but one many companies fudge: There are no "mergers of equals" here; there are acquisitions. "If you won't want to change, don't be acquired," says Norris Woodruff, a general manager in Industrial Systems and an in-house expert on the subject. Sure, GE wants to inspire change rather than compel it, to learn as well as teach, but GE is boss.

A second key point: Integration begins before the deal, with a due-diligence team that's not just financial but includes human resources and general management. They draft a plan to take effect the instant the pens are capped. Says James Conheady, an unsentimental Irishman who runs HR for GE Power Controls, which makes circuit breakers and the like and is based in Barcelona: "You have to move very, very quickly. There's a window of opportunity--an expectation of change--so you must deliver it fast."

First to arrive is someone from finance who is expected, literally on day one, to take control of the books, with the general ledger set up in the format GE uses worldwide. An "integration manager" is a half-step behind--usually a high-potential youngster brevetted into the acquired company. The integration manager doesn't supplant the people running the business. He works full-time on integration; the general manager runs the shop and owns the P&L.

Splitting the roles speeds up integration, and "Do it faster next time " is the conclusion of every postmortem at GE. Many of GE's early European buys--the French medical business (1987), Tungsram (1989), an aircraft-engine repair facility in Wales (1991)--went through two or three general managers before being fully integrated. Leery of Europe's unions and wary of riling host governments, GE had "a certain view that the walk had to be slow, in all the European countries," says Zafirovski, who ran lighting's European operations from 1996 till this summer. Integration plans now run about 100 days. When time's up, basic operational fixes will be under way--like reducing inventory or cutting sourcing costs. Restructuring will be planned, probably begun. Sales forces will have been unified. If the company had a matrix or geographical organization, it will be gone, together with a layer or two of management. It will be replaced by a functional organization--marketing, sales, manufacturing, etc. "Soft-side" initiatives in GE values, expectations, and management methods will have begun.

Most important, leaders will have taken stock of the people to identify "blockers"--who are removed because they impede change--and key people GE wants to keep. To the latter, GE hands out big jobs or glamorous training slots. Says Rino Piazzolla, HR head of Nuovo Pignone: "You have to get rid of blockers and identify high-pots simultaneously. Otherwise you lose the high-pots. You need to show people the future."

There's nothing special about these changes except the speed with which GE does them. But they're worth two to eight percentage points of operating margin, says Joaquim Agut, the Catalonian CEO of GE Power Controls, an $800 million operation composed of ten acquisitions in eight countries, including his family-owned company, sold in 1993.

After generic change comes GEneric change. Agut ticks off a list of GE tools--CAP, CMS, MGPD, bullet trains, QMI, Six Sigma, and more--and says they're worth another eight points of margin. That's an astonishing number, but you can make the case: Ten businesses that today represent about half of GE's European revenues had a combined operating margin of 5% in their first year of GE ownership, vs. an expected 13% margin for 1999.

Says Welch: "These companies have nothing in common except leadership and best practices," but that's not chopped liver. In the view of Merrill Lynch analyst Jeanne Terrile, GE is not so much a collection of businesses as it is "a repository of information and expertise that can be leveraged over a huge installed base." The ability to leverage expertise is why GE's globalization works.

Among GE management tools, several are particularly powerful aids to globalization. One of them, Work-Out, is so fundamental that GE trademarked the term and in its values statement says, "GE leaders...are committed to Work-Out." Work-Outs, begun in 1989, are meetings that can be called by anybody to address any problem, from niggling to humongous, with no boss in the room. When the participants have a plan--kill that stupid form, replace that balky pump--the boss must say yes or no on the spot, no haggling, no waffling. Work-Outs have become so common that there's probably one every day in each sizable GE facility, without management's knowing about it till someone pops in saying "We had a Work-Out and need to talk to you."

Its simplicity is deceptive. Piazzolla says, "We saw 'committed to Work-Out' and asked, 'What's Work-Out? A meeting? Why should we spend time in meetings?'" But Work-Out's few rules totally subvert traditional executive power, and that's radical, especially in Europe.

Says Welch: "Getting a company to be informal is a huge deal, and no one ever talks about it." Informality is the key to GE's in-your-face, just-do-it culture. It allows the other tools to work. QMI, for example, is "quick market intelligence," a tool adapted from Wal-Mart. A regularly scheduled, intense meeting where managers, salespeople, and others pore over data and share anecdotes to get a pulse of the market, QMI can't succeed if people are afraid to tell the boss he's all wet.

In a classic GE twist, the idea that subordinates should push back at their bosses allows bosses to demand more. Magnus Beer, who runs GE's consumer finance business in Scandinavia, vividly recalls his first lesson in "stretch.'' As head of GE Capital's fleet services business in Sweden, leasing vehicles to corporate customers, Beer made his first presentation to his then-boss, Mike Zafirovski. When he was finished, Zafirovski said, "Nice try. You've done a very professional job. Now try something else." Rather than the 13% growth Beer foresaw, Zafirovski challenged him to imagine 25% growth, "then come back to me and tell me what resources you would need to make it happen." That, Beer exclaims, "is a very different discussion!"

GE's European work force has taken to its style with surprising ease. Says Anneliese Monden, who runs GE's leadership- development programs in Europe: "Respect for titles isn't what GE is about, and it's a little strange in the beginning. Once people understand it's safe to think, they love it." According to Kary Wright, an American who is site manager of the aircraft-engine repair facility in Wales, "Personal contact is the key. You need it to share best practices." The plant had belonged to British Airways; it was a cost center servicing BA's engines. GE bought it at a time when the airline wanted cash for new aircraft. That helped GE expand its services business and earned some good will that didn't hurt GE's successful bid later to sell GE-90 engines to BA.

The plant has become Rumpelstiltskin, a cash spinner in a business that's more profitable than selling new engines. Shop-floor workers credit some of the turnaround to GE's wallet--when it was a cost center, management naturally avoided putting money into it--but give equal or greater weight to informality, stretch, and know-how. Says Wayne Glanville, deputy leader of the plant union: "GE's investment speaks for itself. What you can't see is, before, management sat on one side and we sat on the other. Now we have a round table." In the test area, Neil Saxby says, "Teamwork is the key, not authority. All the best ideas come from teams." Almost every GE factory, in every business, uses manufacturing cells where teams share work, rather than assembly lines, which isolate workers. Result: "Information flows so quickly it's almost frightening," says an obviously unintimidated Matthew Gray, who works on Rolls-Royce engines. Color-coded bins and charts--the same around the world--allow workers to monitor stock and see the day's scheduling just by looking at what's in each bin. One area of the plant, which handles accessories such as hydraulic controls and fuel pumps from 70 different suppliers, has just one manager for 130 workers.

Some myths about European labor turn out to be just that, myths. Says William Conaty, GE's senior HR officer: "What I heard six or seven years ago was, You can't restructure; European labor unions are impossible to deal with." Easy they're not, of course. In France, Medical Systems weathered a 55-day strike in 1997. Says European CEO Larry Johnston, who came to France from GE Appliances in Kentucky: "I'd dealt with the union in Louisville, but here it's something else. The three unions are the communists, the socialists, and the anarchists." GE prevailed by being as punctilious about work rules as the strikers were, having a notary record questionable union behavior, storing finished goods at the airport to keep shipments going, and hiding parts on-site so that work went on when deliveries couldn't get through. And by talking: Johnston never let a day go by without meeting with the strikers to tell GE's side of the story.

Clearly it helps that GE is probably harder on management than it is on labor--at Tungsram, for example, some 10,000 people were let go, but about 40% were salaried. Helpful, too, are gestures like one GE made in Wales, where workers had received air-travel perks from British Airways: Unable to do that, GE gave them its highflying stock instead.

"GE is a language--it is a world--and at the beginning it is difficult," says Pier Luigi Ferrara, 62, Nuovo Pignone's chairman. Ferrara knows pressure--he is a famous engineer who designed compressors that deliver ten million pounds of thrust per square inch, 100 times more than the GE 90 engine. He chose to see pressure from GE as a source of new power for a company he loved. Nuovo Pignone didn't understand the value of keeping inventories low, for example; on Welch's first visit there, managers boasted that they never needed to fear a break in production because they had so much extra stock.

More and more, the language of GE is the language of Six Sigma, the quality initiative begun in late 1995. It has become central to GE's ability to operate as a global whole. "Six Sigma" refers to a standard of excellence defined as having no more than 3.4 defects per million--in anything, whether it's manufacturing, billing, or loan processing. GE says it will spend $500 million on Six Sigma projects this year and will get more than $2 billion in benefits. Many are the things quality programs have always delivered--in lighting, for example, reductions in shrinkage (the amount of raw material that doesn't make its way into a lamp) and increases in fill rates (the percentage of a customer's order delivered on schedule), up from 78% in 1996 to 93% today. GE, however, is pushing quality into new realms.

One is asset utilization. Using capital efficiently has always been one of Welch's goals, and Six Sigma gives him new ways to pursue it. For example, GE businesses have long been expected to track productivity gains. (GE measures productivity the right way: Rather than measure output per unit of labor, it measures output per unit of all inputs, whether labor, materials, or capital.) Using Six Sigma data, GE can go further: The company tracks how much of a productivity gain came from new equipment--that is, spending money--vs. from new ideas--that is, from thinking. In Euro lighting, according to Zafirovski, more than half the productivity gains now come from thinking. A big gain: using process improvements to increase output from old plants, thereby avoiding new spending. Says Zafirovski: "The rule of thumb for capacity additions used to be that we needed a dollar of investment to get a dollar of capacity. Now we need 12 1/2 cents." Companywide, Welch watches the ratio of plant and equipment spending to depreciation. It fell from 1.5 in 1996 to 1.2 in 1998; his goal is to get below 0.8. To Welch, this shows Six Sigma's ability to find "hidden factories" built of brainpower.

GE Medical Systems, whose European operations are based near Paris, is leading another advanced Six Sigma effort: using its methodologies to design new products from the customer's point of view. Says Larry Johnston: "Each GE business has taken on a Six Sigma challenge. Design for Six Sigma was ours." GE Medical Europe's contribution: an all-digital mammography system--now in use in Europe but not yet approved in the U.S.--that eliminates film, provides clearer imaging, and allows images to be copied without loss of detail; they can also be manipulated (rotated in 3-D, for example, so surgeons can plan their work better) and shipped over a hospital network.

The new process begins not with prototypes but with a series of Work-Outs with customers. For the mammography system, 16 "luminaries"--top oncologists and surgeons, along with hospital technicians and officials--sat through long sessions with GE experts discussing machine uptime and ergonomics, etc. Customers, in effect, wrote the specifications, working with GE to translate their requirements into technical terms.

A global company needs a common language. Six Sigma is it, more than English. The shared grammar makes it easier to swap ideas around GE. Before, two plant managers, even making the same product, probably had different performance measures. Says Piet van Abeelen, vice president for Six Sigma: "Without Six Sigma, if you run a plant and I run a plant, it's tough to understand your numbers. Then you can say, 'Your ideas won't work, because I'm different.' Well, cry me a river. The commonalities are what matter. If you make the metrics the same, we can talk."

And they talk: Never in business history have so many people talked so much to so many other people as they do at GE. In 1988, when GE started benchmarking other companies and sharing best practices internally, people embraced the notion, but with reluctance, as one might embrace a loved one who had come in from a hard run on a hot day.

It's obvious to anyone who spends time in GE that hesitation has turned to obsession. Says Welch: "This is all about moving intellectual capital--taking ideas and moving them around faster and faster and faster." There is no "GE Europe" per se--no Rome to which all roads lead; the businesses report in to Schenectady, Cincinnati, Milwaukee, Kansas City, etc. But there are dozens of councils: In Europe alone, a Corporate Executive Council for the pooh-bahs, a Marketing Council, Technical Council, Sourcing Council, Finance Council, Human Resources Council, Sales Council, Manufacturing Council, Quality Council, and more. Each business operates similar cross-business or cross-functional networks, which meet for a day or two every few months. Probably every professional in the company--tens of thousands of people--sits on at least one cross-business council. At meetings, everybody is expected to bring something--an idea that made a few bucks, a process that's quicker than the old one. "Go to the European Corporate Executive Council," says Larry Johnston, just named to head it, "and you see 30 people writing all day long."

GE's culture is both deeply competitive and furiously collaborative. It's a company joke that within minutes of Welch's leaving a site, the phone starts ringing as other businesses ask, "What's this thing you told Jack about?" More and more best practices flow both ways across the Atlantic. A system for managing sales forces developed in Barcelona is spreading worldwide. In May 1998, in Florence, Welch heard how a Six Sigma team making turbine packages had cut the total cost 30%. Show me, Welch demanded, and spent an hour grilling the team so he could, when he left, tell everybody else in GE to call Nuovo Pignone.

Ultimately people, not money or tools, make the global GE work. From the beginning, Welch has poured money and energy into creating a massive iceberg of leadership ability. The tip is Crotonville, the company's spiffy executive development campus in New York State; it was a "hovel" when Welch took over, according to one GE executive. What's beneath Crotonville is more important--a passion for leadership development everywhere in the company. The heads of GE's European businesses, asked how much of their time is spent on HR subjects, give answers ranging from 30% to 50%.

There's nothing egalitarian about GE's HR philosophy. It finds the best and culls the rest--period. It can be a cultural shock. Says Tone Rongstand, 32, who has been promoted twice in 1999 and now heads risk management for GE Consumer Finance in Norway: "When I was young I went to a ski competition, and the winner was the one who went in the average time. This is not like GE!"

Leadership development ranks with Work-Out, Six Sigma, and idea swapping as GE's best tools for globalization. It starts with new hires--or people at newly acquired companies. Engineers might get into TLP, a tough, two-year technical leadership program that involves three eight-month projects interwoven with classroom work in project management, process-improvement methods, and so on. Similar courses exist in every functional discipline from finance to information management, as well as Six Sigma training leading to "black belt" and "master black belt" status. Still more courses are offered at intermediate and advanced levels, always with lots of time in the field.

The training unleashes a number of forces, of which content is least important. It creates networks that unify the company and leverage its diversity. It's impossible to exploit GE's size if people don't know one another. Second, lavish investment in high-pots sparks ambition. Says Danielli Schenarelli, a sourcing expert at Nuovo Pignone: "The American idea of the self-made man translates into a management style of self-selected leadership." When someone at GE feels the urge to lead, he just starts leading, regardless of age or title--and has almost infinite running room. Raf Jansen, an electrical engineer, joined GE from college in 1994. In March 1998, Joaquim Agut spotted him and asked him to lead Power Controls operations in Poland. The project involved converting a joint venture into a 100% GE-owned operation, building (in six months) a greenfield factory in Klodzko, and moving production of small circuit breakers there from high-wage Western European countries. Jansen is 30 years old.

Sophisticated HR systems make sure that no sharp knife stays in a drawer. These include annual self-evaluations and feedback as well as a yearly, companywide review of talent. Welch himself reviews some 3,000 people. Top managers pursue top people like kids chasing an ice cream truck. At the Hotel Gellert in Budapest one spring morning, Mike Zafirovski bought breakfast for a 30-year-old hot-shot high-pot from Medical Systems, whom he wanted for lighting's operations in Istanbul. Said Zafirovski: "Larry Johnston wants to keep him. He has sexier technology, but I can offer P&L responsibility." (Zafirovski won.) Leaders also compete to see who shares talent best. When Nils Albert, who runs a leasing business for GE Capital in England, needed a salesperson in Spain, he called Agut in Barcelona, who leaped at the chance to give him one. Says Agut: "Now everyone in Europe knows that Nils got someone from Power Controls, so we get a reputation for developing people--which helps me."

Leadership Darwinism cuts both ways. Since the beginning of 1999, GE has been pushing a harsh forced ranking of employees. For years, GE has divided salaried employees into A, B, and C players--with pay and promotions disproportionately given to A players, and C players left back. Now an "organizational vitality" measurement requires every manager to rank his staff--even if they're all A players--into a "top" ten, "strong" 15, "highly valued" 50, "borderline" 15, and "least effective" ten. "It's wild," says Welch: "You send your top ten on, and see how many of them get into the top ten for the whole business. Or you look at what functions are represented. If there are no tens or 15s from finance, it means the whole function is stuck in the middle." It's not pretty, but Welch thinks a merciless push to upgrade human capital is vital.

There are, Welch says, three stages of globalization. First is globalization of markets--acquiring the assets to sell in any market you choose. Second is globalization of sources--becoming able to buy or build wherever it's most advantageous.

Welch sees GE entering phase three: "Globalizing the intellect of the company"--going wherever the best brains are. Says Welch: "It means using Russian engineering and Indian software--not to arbitrage labor costs, but because these are the best people you can find." GE gives stock options as generously to Europeans as it does to Americans. As a result, there are two dozen dollar millionaires at GE Medical in Europe. Non-Americans fill Crotonville's top classes in rough proportion to their numbers in GE's total employee base. The head of GE's corporate audit staff is French, Six Sigma's leader is Dutch, and now a European leads one of GE's core businesses. But, Welch says, to globalize intellect deep into the organization, "is the hardest thing a company can do, because no one is safe then."