(FORTUNE Magazine) – After running IBM for more than a year and a half, CEO Lou Gerstner has revealed himself to be something other than the revolutionary whom the directors of this battered and demoralized enterprise once seemed to want. They first tried to recruit Jack Welch of GE and AlliedSignal's Larry Bossidy, both in-your-face practitioners of radical change. What's becoming clear is how fundamentally Gerstner's approach to leadership differs from theirs. Dark, compact, and vividly intelligent, Gerstner describes himself as a "general manager." Expected to lead the biggest corporate transformation of all time, he instead seems to be attempting a conventional turnaround: deep-cleaning and redecorating the house rather than gutting and renovating it. Results so far have been surprisingly good -- profit margins have stabilized and the stock has soared -- so why complain? Only because Gerstner's cautious methods increase IBM's already unnerving exposure to marketplace risk. The CEO deserves full credit for cutting costs and introducing a new sense of urgency, but a big factor in IBM's return to profitability has been luck. Unexpectedly high demand for mainframe computers has given the company temporary respite from the inevitable shift to less lucrative products. No one can predict how long the grace period will last. Gerstner's IBM is more or less the same institution that has vaporized $60 billion of market value since 1987. Still hobbled by high costs, infernal complexity, and a neurotic corporate culture, the company lags behind Compaq Computer, Hewlett-Packard, and other leaders in the fast-changing, intensely competitive computer industry. Weakness invites attack: In August, for instance, PC-market leader Compaq launched a price war certain to bloody IBM. Once admired as the world's greatest corporation, IBM simply is not competitive anymore -- and Gerstner concedes that it may not become competitive for another couple of years. If his luck sours, that could be too long to wait. One measure of the difference between IBM's CEO and a revolutionary is his sense of time. Gerstner readily acknowledges the uncertainty of the company's fate. But having determined that IBM must reduce its expenses by $8 billion to match competitors, he is proud that his team has cut $4.8 billion, or about 20% of total expenses, so far. "Our results on expense reduction have been more than satisfactory," he says. "Transforming IBM is not something we can do in one or two years. The better we are at fixing some of the short-term things, the more time we have to deal with the long-term issues. You can't really start addressing the long-term issues unless you've got a stabilization." A revolutionary might disagree, focusing instead on what it takes to win. "Maybe he needs to throw a hand grenade around the building to wake people up," says one who thinks Gerstner is moving too slowly. Leaders who insist on getting results right away usually break more glass than Gerstner has. Of the 37 executives in Gerstner's Worldwide Management Council, only eight (including Gerstner) come from outside IBM. When Compaq faced its own crisis a few years ago, CEO Eckhard Pfeiffer attacked the corporate culture head on, abandoning Compaq's quality-at-any-price business model and cutting expenses by 30% in just nine months. Result: the world's best PC company. Gerstner's pursuit of long-term goals requires keeping investors happy quarter to quarter. At a recent price of $68let's update all these numbers on closing day per share, IBM's stock has risen 32% during the CEO's first 17 months with the company. By comparison, AlliedSignal stock rose 73% during Bossidy's first 17 months. In fairness, Gerstner's challenge is tougher than those faced by the leaders of GE, AlliedSignal, or Compaq. IBM is bigger, more complicated, and fundamentally weaker than they were. So far, though, his accomplishments fall short of the standard set by those companies' CEOs. Gerstner has never led a radical transformation himself. Nor does he claim great expertise in technology. A star consultant at McKinsey, Gerstner rose to No. 2 at American Express, where he ran the charge card division, then became CEO of cigarette- and cookie-maker RJR/Nabisco. His record is mixed. At American Express he produced handsome increases in divisional profits, but the business he left behind is troubled today. At RJR/Nabisco, which Kohlberg Kravis Roberts had purchased in a leveraged buyout, he cut debt by $12 billion but couldn't match other tobacco companies' increases in market value. The IBM directors who finally settled on Gerstner were publicly humiliated and presumably exhausted after presiding over the company's fall from grace and the ouster of chief executive John Akers. They awarded Gerstner a nine- year contract enriched with a $4 million signing bonus, options on 500,000 shares, and compensation that added up to $7.7 million last year. Since then, nine of 18 board members have stepped down, including three of the seven members of th search committee. The troubles Gerstner faces are not of his own making. Drugged by seemingly limitless profits from mainframe computers, IBMers from the boardroom on down failed to respond as the industry's center of gravity shifted to PCs during the mid-1980s. As the financial charts at right make painfully clear, Big Blue performed like a ballistic shell, reaching the peak of its trajectory just as its momentum gave out, then plummeting. Since 1990, IBM's gross profit margin (sales minus cost of goods sold) has dived a sickening 17 percentage points, to less than 40% of sales. Last year IBM's gross profits were $14 billion less than in 1990 -- a difference equivalent to the total sales of 3M, one of the largest industrial companies in America. The plunge in profitability is what's forcing IBM to drastically change the way it operates. Since late 1992 it has managed to hold gross profit margins steady, partly by eliminating jobs. Starting before Gerstner arrived, IBM has cut 170,000 jobs worldwide -- almost as many as GE has lost under Welch. (Cutting workers in Europe is remarkably difficult -- see box, page 88.) Last year, IBM reported an $8.1 billion loss due to restructuring charges but broke even on operations; for the first half of this year net income totaled $1.1 billion. Gerstner candidly acknowledges danger ahead: "Margins will be under pressure for the rest of the decade." Even some of his harshest critics greatly respect Gerstner's abilities as strategist and manager, and the pace of change has accelerated in recent months as his emphasis has shifted from study to action (see time line). But the body of opinion now starting to jell holds that his leadership may be too deliberate and cerebral -- too McKinseyish -- to transform IBM's self- defeating business practices and inward-looking corporate culture in time. For benchmarks, FORTUNE looked to GE, AlliedSignal, Compaq, Tenneco, and the few other large corporations that are models of the dramatic reengineering IBM hopes to achieve. In addition to interviews with a dozen top IBM executives, FORTUNE talked to security analysts, computer industry specialists, competitors, former employees, and a Greek chorus of CEOs and other well- informed observers, many of whom spoke on condition of anonymity. John Naisbitt, author of Megatrends and Global Paradox, sums up what many of the critics think: "Gerstner is very, very good at what doesn't work anymore." Adds University of Michigan professor Noel Tichy, who co-authored a book with me about the Welch revolution at GE and who has since consulted for Gerstner's IBM: "Lou hasn't made any big mistakes, but he's still putting the base camp together at the bottom of Mount Everest. In an industry moving at the speed of light, there's very little time for that." Critics fault the CEO for being an incrementalist who relies mainly on longtime IBMers and is not fundamentally challenging the IBM way of doing business. In person, Gerstner radiates intensity. A fleshy, impeccably groomed man with thick black hair, hazel eyes, and smooth round cheeks, he is a tightly coiled package of high-voltage brainpower. Like some actors and TV personalities, he has the gift of appearing more in focus, more brightly lit, than anyone else in a room. Gerstner can be impatient, rude, and abrupt, but his intellectual honesty is striking and he is quick to acknowledge a valid point even when it is critical. Above all, the guy is a brain, blessed with a crisp, lucid, orderly intelligence that sorts and synthesizes ideas at remarkable speed. That data-processing power helps make him an unusually effective communicator, capable of distilling large volumes of complex information into memorably simple ideas and phrases. Consider this blunt exchange, from a question-and-answer session at an IBM plant: Employee: "One of your key things, Lou, seems to be getting closer to the customer. Frequently we find that we need to trade off customer requirements in favor of strategic decisions -- platform-type support. Can you make specific recommendations for how we can achieve a successful balance?" Gerstner: "I don't know what 'platform support' is, but let me answer your question very simply. If there's a key customer requirement, do it. Whatever this thing is -- 'platform whatever' -- get it out of the way." The CEO is a perfect model of the intense urgency he wants all IBMers to bring to their jobs. He can outwork practically anybody. According to strategy chief Jim Cannavino, the irreverent, self-educated executive who built IBM's PC business, Gerstner "reads a stack of stuff that is probably half his size every day." Gerstner packs his calendar with visits to IBM locations and customers around the world. He is so unendingly busy that John Thompson, the executive responsible for mainframes, minicomputers, and a bulging portfolio of other products, recently had to wait a month for a routine meeting with him. IBM has just as much personality as its boss. Its culture was shaped by decades of rapid growth and gross profit margins that long hovered around 70% of sales in its mainframe businesses. World-straddling IBM represented the creme de la creme of business, hiring the smartest people, teaching them how to dress, and conferring upon them, as if by divine right, the power to crush competitors. To resolve conflicts between staffers and line executives, the late Tom Watson Jr., son of IBM's founder and its most influential CEO, instituted a process that permitted anyone who disagreed with a decision to "non-concur." The conflict would then go up the hierarchy for resolution. As long as Watson and his successors pressured IBMers whenever decision-making bogged down, the system worked. Over time, though, it degenerated. In a company whose work force peaked at 406,000 in 1985, the system came to encourage infighting and discussion to the point of paralysis and beyond. Lowest-common-denominator thinking often prevailed. As executives sought the autonomy they needed to get things done, much of IBM devolved into a collection of fiefdoms based on product, job function, or geographic location. By the time Gerstner arrived, remarks an IBM hand, "they were all shooting at each other, not at the enemy." As everyone knows, the catastrophic turn in IBM's fortunes came with the popularization of the personal computer. Long the leading producer of PCs, IBM had access to all the information it needed to capitalize on the trend. Indeed, its strategy wizards created some plans that Gerstner, who has reviewed planning documents dating back as far as the late 1970s, regards as brilliant. What went wrong? "IBM didn't implement those plans," he says dryly. A stroll through the broad corridors of IBM's headquarters in leafy Armonk, New York, helps explain why. Built in the early 1960s, the buildings are still carpeted in alarming shades of go-go orange that bring to mind bell-bottom trousers, the Watusi, price/earnings multiples of 40-plus, and other emblems of the era in which IBM reigned supreme. When the world beyond those corridors changed, IBM non-concurred, and the rest is history. Gerstner has made it plain that the days of writing elegant strategies and then filing them for posterity are gone. To force action, he frequently arbitrates disputes with snap decisions and grants sweeping authority to individuals or small committees. He has launched eight enormously broad reengineering projects, each "owned" by a single executive. Topics include hardware development, production processes, and customer fulfillment. For example, in addition to her responsibilities as a group executive, Ellen Hancock is charged with reengineering IBM's software development -- not easy, since six different divisions produce software products. Gerstner has already greatly enhanced the quality of thought at IBM. Before he arrived, IBMers documented their presentations at meetings mainly with slides showing simple bullet points. Decision-making suffered as a result. Says Jim McGroddy, chief of research: "The spoken word was fleeting, and therefore not inspected as thoroughly as it should be. And the written word was very incomplete, and therefore not inspectable." By contrast, Gerstner requires arguments fleshed out in writing. Distributed in advance, these documents are required reading before meetings begin. The result, says McGroddy, is that "the time you spend together is used to move beyond what's already been written down."

GERSTNER'S most visible intellectual achievement is the distillation of hundreds of IBM goals into a simple list of strategic imperatives. Although in his first days as CEO he remarked, foolishly, that "the last thing IBM needs right now is a vision," this list constitutes a lucid expression of where he wants the company to go. Key points include a belief that IBM's size, scale, and technology can once again become sources of competitive advantage. Regarding details about strategy as competitive information, the CEO is reluctant to elaborate much on how those imperatives will play out in IBM's diverse businesses. This company has more strategic projects than a porcupine has quills. At the corporate level, for instance, IBM is consolidating all its advertising -- some $500 million of annual billings -- at Ogilvy & Mather Worldwide. Hoping to establish itself as a leader in computer networking, IBM is introducing a broad line of products that use a state-of-the-art technology called asynchronous transfer mode. For all that the company has in the works, worry about IBM under Gerstner boils down to four main concerns: threatened revenues, high costs, unwieldy organization, and excessive reliance on the CEO himself. The old businesses that produce most of IBM's revenue and profit -- big iron plus related software and services -- probably will continue to shrink, and new businesses may not replace them. "If I were Gerstner, I wouldn't be too happy," says a CEO whose company competes with IBM. "In two years IBM's revenue is going to start to go down -- I expect a good 20% or 25% whack." Gerstner, by contrast, says he sees no reason why revenues can't grow. Fair enough, although advances in technology will make sales increases dauntingly hard to achieve. IBM recently has sold record amounts of mainframe processing power, measured by MIPS (millions of instructions per second), and has orders in hand for more mainframes than it can readily build. But the price IBM gets per MIPS has fallen by 70% in the past two years, resulting in lower revenues. The mainframe business is still very profitable, but sooner or later -- observers say in five years at the outside -- it will wither as most customers find the courage to shift to so-called client-server systems, which are networks of smaller computers. When that happens, other IBM businesses that depend on big iron will also decline. Most of IBM's $10.9 billion of annual software revenues, for example, are from programs used on mainframes. Some of Gerstner's key strategies for growth, though well reasoned, are fraught with risk: In the face of able competition, IBM is banking on attracting millions of customers to a new, unified product line based on its Power microprocessors and Workplace software. Power is IBM's name for the family of chips that provide the smarts in many of its popular RS/6000 engineering workstations. Based on RISC technology (techspeak for reduced instruction-set computing), the chips provide substantially more bang for the buck than the older technology used in Intel's Pentium or 486 chips. By packing more processing power onto a given amount of space, RISC designs enable manufacturers to produce more chips from each silicon wafer, which lowers costs. Gerstner's plan is to leverage that cost advantage by using Power chips in most of the computers IBM makes, from PCs to small mainframes. Standardizing ! parts should benefit customers by making possible lower prices and enabling different types of IBM computers to run the same software. The Power strategy is designed to take away chip revenues from Intel. That is crucial, since ever more of the economic value in computer systems is concentrated in the silicon: more than 25% of total value today, up from 15% during the 1970s by IBM's reckoning. And unifying the product line could heighten the appeal of one-stop shopping at IBM. Whether IBM succeeds with Power will depend on a parallel strategy in software.Called Workplace, this software would mediate between the computer hardware and application programs, such as word processors or databases, and save customers money by making programming easier. But the technical challenges of implementing the Power and Workplace strategy are so formidable that some industry watchers wonder if IBM's technologists have bamboozled Gerstner. "It's nonsense to think they can have one common software," warns one. "That strategy is completely unexecutable." Cost cutting under Gerstner -- drastic though it may seem to employees and widely praised though it has been on Wall Street -- looks like too little, too late. From the first, efforts to analyze IBM's cost structure were hampered by the company's decentralized and embarrassingly ineffective information systems. Jerome York, the chief financial officer Gerstner hired from Chrysler, explains: "Information on any specific country is reasonably easy to get, but when you require information that cuts across IBM in its entirety, you find that the information systems in these various countries are not compatible." Determining the company's global advertising expenditures, for instance, literally would require sending people to file cabinets around the world to check vouchers. The company has launched a major reengineering of its information systems, consolidating 159 data centers into a dozen or so, while reducing the annual cost of such systems by $2 billion. Meantime, York has focused on cutting the expenses IBM can track. After studying data on other computer companies, he determined that IBM needed to chop $7 billion from annual expenses to be competitive; later he raised the goal to $8 billion. Under Gerstner, IBM's expenses have dropped from 40% of sales to 33%, still four percentage points higher than those of a weighted average of competitors. The slow pace is more worrisome because of York's somewhat misleading semantics. When he talks about "expenses," he means only sales, general, and administrative expenses plus R&D, not IBM's $39-billion-a-year cost of goods sold, a much larger category that includes manufacturing and raw-material costs. IBM won't disclose its reduction targets for these, but York says he expects them to drop to competitive levels by the end of 1996. In other words, IBM will not be cost competitive for another two years -- an eternity in the computer industry. One of Gerstner's greatest challenges is finding a workable organizational structure for IBM. His sensible-sounding theory is that IBM's competitive advantage lies in its potential to offer one-stop shopping to customers seeking solutions to their technology problems. Before he arrived, Big Blue was moving toward a breakup into 13 independent units: one for mainframes, another for PCs, a third for disk drives, and so on. That decentralized structure appealed to the units' leaders, who craved autonomy. Many outside observers agreed, figuring that focused businesses would compete best against smaller, more agile companies like Microsoft, which devote their full resources to a particular specialty. To gain advantage from a unified IBM -- which offers almost every imaginable computer product, from software and silicon chips to services and supercomputers -- Gerstner must find ways to knit 235,000 employees into a single, efficient team. No other company has ever faced an organizational challenge of this magnitude. Even GE, which has grown to roughly the same size as IBM, simplified the problem by subdividing itself into a dozen free-standing units, each devoted to a product line such as turbines, light bulbs, or financial services. One way to make the most of IBM would be a matrix structure, in which individuals may be answerable to two or more chains of command. But by involving more people in a routine decision, matrix structures can inhibit action and muddy accountability. Truly unifying IBM's 159 profit-and-loss units would result in an organization chart so complex it would have to be shaped like a sphere -- or a black hole. Digital Equipment, the No. 3 computer maker, decided to abandon its matrix structure this summer. CEO Bob Palmer explains that endless confusion about who was responsible for what overwhelmed the advantage of having one sales force pitching all of Digital's products. Gerstner counters that IBM doesn't need the perfect organization. "Eighty percent of the time people know what they're supposed to do without needing someone else's input," he says. "Increasing teamwork is going to be complicated, but I'm not too worried. If the organization doesn't work right one way, we'll change it." Indeed, at American Express, Gerstner reorganized almost annually.

THE ENERGETIC Gerstner's main failing is that he still seems to be driving IBM's transformation almost single-handedly. It's hard to imagine who would take up the reins if he were incapacitated. Smart and committed though all the senior officers appear to be, not one has been through the sort of process Gerstner is trying to lead. While he indoctrinates this cadre, the CEO must rely on his own personal power to force change. "Lou has his hand firmly on the throttle," says John Thompson. "We are not at the stage yet where you stand back and let the organization take over and implement what you have set up." But any sense of "l'etat, c'est moi" could impede the transfer of power that seems to characterize successful transformations: from the CEO as the main driver of change to momentum sustained through pumped-up employees. Gerstner acknowledges the issue: "I'm not unhappy with the ability of the organization to implement the things that I have personally been involved in and have been leading," he says. "There are scores of people who are already on the program. But at some point, we will need 20,000. And we're far from 20,000 now." Gerstner may not need quite that many committed helpers to bring IBM's transformation up to speed -- but he needs more than he has. CEOs trying to change the behavior of large numbers of employees usually rely heavily on the human resources department, which controls the essential levers of appraisal and compensation. IBM's is in disarray. Early on, Gerstner hired Gerald Czarnecki, then head of BankAmerica's Hawaii unit, to head HR. He didn't last long and hasn't been replaced. That's a shame, since IBM's outdated HR systems are impediments to change. Gerstner has dramatically altered the financial incentives for top executives, basing roughly 75% of their variable pay on the overall performance of the company, vs. about 25% before. But he says he relies on line managers to set the proper incentives for their people. Perhaps his trust is misplaced. For most of IBM's sales reps, variable pay amounts to less than 10% of their compensation. Asked whether his line people know how to set the right - incentives, Gerstner laughed and replied, "No. That's why we need a new HR person." Having bought time by stabilizing IBM's finances, Gerstner is hoping to transform the company's culture over the next five years. As shareholders and directors ought to know, he will be the luckiest CEO in history if the brutal high-technology marketplace allows him that long. The stakes could not be higher: The crushing force of economics probably will give IBM just one more roll of the dice. If Lou Gerstner fails, it may be too late for another CEO to make a winner of IBM.

BOX:

LONELY AT THE TOP? Gerstner wants 20,000 allies to remake IBM.

CHART: NOT AVAILABLE CREDIT: NAI LEE LUM FOR FORTUNE/SOURCES: IBM; SMITH BARNEY CAPTION: THE NUMBERS: better but not good

CHART: NOT AVAILABLE CREDIT: GRAPHICS BY NAI LEE LUM CAPTION: GERSTNER'S FIRST 17 MONTHS