BREAKING UP IBM Facing horrendous challenges, John Akers is taking Big Blue apart and may even sell stock in the pieces. His reforms rate a B- so far, but the job will need an A+ effort.
By David Kirkpatrick REPORTER ASSOCIATE Jennifer Reese

(FORTUNE Magazine) – THINK YOUR company has troubles? IBM faces practically every challenge known to management. Chairman John Akers is forcing perhaps the most complex company in the world onto a course of vast, painful change. His goal: to uproot a structure and culture formed when IBM had no serious competitors. The reforms he announced in December are more far-reaching than most outsiders have recognized. But is he pressing the message hard enough to permeate this immense organization? Probably not. Besides its bulging, lethargic bureaucracy, IBM's difficulties include overdependence on high-margin mainframes, when computing power has become a desktop commodity; commitment to proprietary hardware and software, when customers insist increasingly on open systems compatible with other manufacturers' products; emphasis on hardware, when software and services are ever more important; and an inability to get quickly to market with the new machines that periodically remake the industry (see chart, page 47). For all those reasons, the numbers look terrible. Last year was the first since 1946 that Big Blue didn't get bigger. Revenues were down more than 5%, to $65.4 billion. After huge special charges, chiefly $3.4 billion related to restructuring, the company lost $2.8 billion -- its first deficit in history. A year ago many of IBM's horde of critics -- investors, analysts, and consultants -- huzzahed when Akers angrily warned a group of IBM managers: ''The business is in crisis.'' Finally, the company was acknowledging the obvious. But in a FORTUNE interview, his first extensive talk with a U.S. journalist since the December restructuring, Akers made clear that he is no longer in an alarmist mood. He believes the retuned organization will, in time, get with the program and thrive again. The company should be back in the black this year and will even pick up mainframe market share. Akers blames much of IBM's difficulty on the computer industry's anemia. He argues: ''The problem has been adjusting to change, and to a market that grew at 15% and now isn't growing at all. That's a big, big difference.'' Certainly. But that ''problem'' hasn't stopped Apple Computer, Hewlett- Packard, Sun Microsystems, or even troubled Digital Equipment. Their revenues climbed 7% to 31% last year. Already on the hot seat, Akers has turned up the thermostat by setting specific near-term performance targets: For return on stockholders' equity, which has languished at or below 15% since 1986, he promises 18% within two or three years. Sales, general, and administrative expenses, now a fat 33% of revenues and a figure Wall Streeters complain is way too high, are to plunge to 23% by 1994. He targets sales growth of 7% to 10% per year, and return on assets of 8% (vs. 6.9% in 1990). Free cash flow, a mere $500 million last year, should hit $4 billion by 1994. IBM's sheer vastness makes it tough for Akers to change the culture in a hurry. Big Blue is still more than four times the size of Digital, the No. 2 computer maker, and ten times larger than Apple. It sells five times more software than Microsoft. Many of the consequences of its size and past success are startlingly similar to those of another humbled American corporate behemoth, General Motors (see box, page 49). Some of Akers's reforms have been under way for years; others are new. He is wrestling to push power downward in what was for decades the consummate top- down company. He's breaking Big Blue into 13 divisions and giving them increasing autonomy -- to the point that IBM may soon distribute shares in some of these new Baby Blues to stockholders. He's struggling to expand the pivotal services business and make the company far more responsive to customer needs. He's continuing an overarching five-year effort to cut costs, so far with uneven success. Akers's place in history rides on making his ambitious program work. If he follows IBM tradition, he will retire once he turns 60 in December 1994. That would leave him just 2 1/2 years to ensure that his tenure has not been a failure. ''We are now on record as a team with a financial objective and a period during which we intend to get there,'' he says. ''So we are now publicly accountable.'' For a preliminary accounting of the wisdom and potential success of IBM's self-proclaimed transformation, FORTUNE spent two months talking to more than 65 analysts, consultants, customers, employees, and leaders of competitors from Lester Alberthal Jr. (EDS) to Joseph Zemke (Amdahl). Besides Akers, a score of IBM's top executives, including the general managers of eight of the 13 component businesses, weighed in. Based on that reporting, FORTUNE herewith grades Big Blue on its eight key strategies, using two criteria: how sweeping the reforms are and how urgently the company is pursuing them. The best mark: A-, for cutting staff while maintaining its cherished no-layoff policy. The worst: D, in remaking the corporate culture. IBM's overall grade point average: B-. That puts the world's biggest computer company far from the head of the class. FORTUNE's report card on Big Blue, strategy by strategy:

GIVING MORE AUTONOMY TO THE BUSINESSES B+ Says Todd Hixon, electronics industry specialist at the Boston Consulting Group: ''The whole IBM business model has been based on centralization -- in mainframe computers and in corporate organization.'' By dividing the mammoth company into 13 distinct parts, Akers hopes to reinvigorate it by creating leaner, more manageable units that he says will grow still more distinct over time. Nine are ''manufacturing and development lines of business'' (LOBs for short, pronounced as in tennis); they produce hardware and software. The other four are geographically organized ''marketing and services'' or M&S companies, which sell what the LOBs make. Each member of the new ''federation of companies'' is now essentially autonomous, says Akers. The LOBs function as manufacturers and wholesalers, the M&S companies as retailers. Akers seems to want IBM to become something like one of Japan's keiretsu, groups of companies with common goals but substantial independence to seek out and exploit business opportunities individually. This spring, for the first time, each Baby Blue general manager signed a contract with the corporate management committee agreeing to targets in seven areas: revenue growth, profit, return on assets, cash flow, customer satisfaction, quality, and employee morale. If he or she exceeds the targets, he and his employees get special bonuses. If a general manager consistently falls short, he could be a goner. The general managers love their liberation. Says Robert LaBant, head of IBM North America: ''This has unleashed an unbelievable amount of energy and creativity like I've never seen before. It's exciting. It leaves me breathless.'' Adds the more restrained Renato Riverso, who runs IBM Europe/ Middle East/Africa, the largest M&S company: ''These are radical changes, not just a readjustment of company structures. It is changing the company from the base.'' Part of Akers's intention seems Darwinian. In the past, IBM managed itself for the good of the whole; less-than-stellar businesses and inefficient practices were allowed to survive. Akers now hopes that internal competition and market-based transaction pricing between the businesses will help determine which succeed and which fail. The financial results for all 13 Baby Blues will probably be made public by early 1994, some possibly earlier. To attune the LOBs to market realities -- as well as to make extra money -- Akers is encouraging them to sell their products on a larger scale to other companies in the industry. That business will help determine how much the LOBs should charge one another; IBM expects it to amount to $1 billion in 1992. The target for next year: $3 billion. Even Joe Zemke, CEO of Amdahl, a major competitor on IBM's cherished mainframe turf, says he is now in the midst of promising talks with Big Blue about acquiring IBM technology. Perhaps the most gut-wrenching reality confronting Akers is the inevitable competition among parts of his own company. Smaller, cheaper machines developed by the Personal Systems LOB -- personal computers and workstations, often configured in networks -- are rapidly cannibalizing their more expensive cousins. Two IBM midrange computers, nearly identical in speed, memory, storage, and software but produced by different LOBs, compete head-on: the open-systems RS/6000 workstation and the AS/400 minicomputer, with predominantly closed proprietary software. Computer Economics Report of Carlsbad, California, recently put the total cost of buying and running the heavyweight AS/400 for two years at over $1 million, while the fast-stepping welterweight RS/6000 came in at $218,000 comparably equipped -- barely one- fifth as much. Peculiar, no? James Cannavino, the blunt-talking general manager of the Personal Systems LOB, which makes the RS/6000, isn't afraid to talk about head-to-head competition with his peers. ''The reality now is that there is going to be product overlap,'' he says. ''That may mean we are going after business that Nick ((Donofrio, who runs the mainframe LOB)) wants. I'm not going to make my products worse so Nick wins. But I'm not going to feel bad if Nick wins, and I'll feel terrible if Hewlett-Packard wins. I'm going to make whatever is competitive in the industry, period.'' Sure enough, at a recent briefing for consultants and reporters, Nobuo Mii, a Cannavino deputy, spoke unequivocally of his group's intention to ''bring the technology of the mainframe to the desktop.'' Personal Systems exemplifies the new autonomy vividly, because Cannavino has been among the most aggressive LOB general managers. His people are starting to take cues from their loquacious boss and broach subjects IBMers never were supposed to talk about before -- like problems they have dealing with other parts of IBM. Cannavino, 47, widely mentioned as a potential successor to Akers, is the only top official in the company without a college degree. One story has him temporarily stealing a multimillion-dollar mainframe to get a project jump-started when he was still an engineer. His enthusiasm for restructuring may have a personal resonance: This onetime overweight smoker has quit and shed some pounds himself. Truly independent businesses have independent ownership. IBM's 770,000 stockholders may soon find themselves with a number of shares where they had only one before. Says Akers: ''Our intention is not to continue to be the 100% owner of each of these businesses.'' Treasurer Robert Ripp spells it out in a presentation he has been making lately to big institutional stockholders: ''I would like to find a stock mechanism so investors can see where our capital investments are being rewarded and the market can see where we're being successful.'' He insists there's no need for new equity to raise cash. The model Ripp calls ''interesting'' is USX, which last year split each share into two parts -- one for its energy operations and one for steel, giving shareholders a pure play in each. He says such a change at IBM could not happen before the end of 1993 because Securities and Exchange Commission rules require three years of audited income statements from any potential spinoff. But he leaves the strong impression that by then a reasonable person might expect to see shareholders get stock in at least a couple of the businesses. Says Akers: ''The value of IBM may very well be a lot more in the sum of the parts than the whole.'' The example of AT&T suggests he's right (see box, page 52). Many students of IBM remain to be convinced that decentralization and intramural competition will improve performance. Among them is the investment manager for an institution that owns several million IBM shares, who says, ''The pluses are that they may be able to accelerate the growth in revenue if all the divisions are able to sell whatever products they want. But the risk is that as they release all these divisions to compete with one another, gross profit margins will deteriorate even more than they have.'' For his part, Akers foresees a careful juggling: ''To take a current problem, I doubt that the optimization of worldwide manufacturing capacity for nine manufacturing and development businesses, one by one, would be the right optimization for the IBM company. We're having some wonderful debates about this manufacturing stuff now.'' For example, Akers says a LOB might prefer to stop manufacturing in a particular country -- but the country marketing manager could need to keep the plant going for political and commercial reasons. Splitting IBM into more manageable pieces is the right course. The old centralist mentality still exerts a powerful pull, as anybody who spends time with employees at various levels around the company soon discovers. The urge to ''defer to corporate'' suits the IBMer's soul the way peanut butter suits jelly. Akers must push authority and responsibility much lower in the organization for the average employee to get the message. He'd better keep talking crisis.

REMAKING THE CULTURE OF COMPLACENCY D ''We have all the habits that generated one of the most successful enterprises in business history,'' says Akers. ''We also have all of the habits of a business that hasn't performed up to our expectations.'' His understatement is telling. The company is deeply complacent. If any one factor could derail his efforts at reform, this is it, and he seems insufficiently concerned. The company is taking many steps, some mainly symbolic, to convey to all and sundry that life is not as it was. Last fall Akers removed George Conrades from the helm of IBM U.S. Says a director: ''Akers shot somebody in public, and he did it to send a message. George wasn't performing.'' This attitude is spreading downward. Says Dennie Welsh, who runs a subsidiary that operates computer systems for U.S. customers: ''If the management team doesn't perform, ^ there will be a new management team. That in itself is a little new for IBM.'' Akers has encouraged the LOBs to bring in at least one outsider at a high level this year, according to Sam Albert, a consultant who retired from IBM in 1989 and retains good sources inside the company. Nick Donofrio says he expects to add as many as three to his mainframe group over the next year. According to many of those who deal with IBM, the company remains extraordinarily mainframe-centered. Even within Jim Cannavino's LOB, the group that makes PCs is called the Entry Systems Division. ''They still think of the PC as just another portal for people to use their mainframes, which have been their cash cow for decades,'' says the CEO of a PC industry company who works closely with IBM. Like many, he is afraid to take the risk of publicly excoriating Big Blue. How about the famous IBM blue suits, white shirts, and dark ties? Cannavino has taken to wearing casual attire in public. At the announcement of his new OS/2 operating system at the Chicago Comdex trade show in April, he wore a yellow V-neck sweater with an IBM logo on the chest. Two other top executives also went tieless. Morale remains in the pits. Many employees feel for the first time that their jobs are in jeopardy. They see the company being tougher on them and their friends and at the same time worry that somehow their managers are killing off the company. Says an executive at a software firm who spends a lot of time with IBMers: ''The people are lost. After years of being part of the IBM family, it's as if somebody shot the momma and the poppa.'' One of Akers's best innovations has been to move toward paying people according to how well their part of the business does. While fewer than 100 executives outside marketing were on incentive pay plans several years ago, today 1,500 are. This year Akers spread $200 million among a number of business units to help motivate workers at all levels. All U.S. employees can now add as much as 3% to their pay if their unit exceeds its targets, and some overseas businesses have similar plans. That may sound paltry, but for IBM it's revolutionary.

Reshaping IBM's petrified culture is like reshaping the Rocky Mountains, and so far Akers has wielded only a pickax. Despite all his efforts, some IBMers still don't know which LOB they're in. Some, asked what they think of the restructuring, respond, ''Which one?''

RETAINING A SINGLE SALES FORCE B- This may be IBM's most complex problem -- how to give individual businesses the benefit of self-management while presenting a single face to the customer. The strategy so far: Maintain a monolithic worldwide sales organization, divided up geographically. IBM's critics fiercely debate this approach. Even some of the severest think Akers has chosen correctly. Says Stephen Cohen, who follows IBM for Soundview Financial Group, a brokerage firm in Stamford, Connecticut: ''To totally break up the sales force to the LOBs would risk confusing customers and inviting competitors to make hay.'' In the late 1970s the company tried separate sales forces for large systems, small systems, and office products like typewriters. Customers went crazy getting calls from three different salesmen. Others disagree with Akers. James Moore of GeoPartners, a Cambridge, Massachusetts, consulting firm that specializes in large high-tech companies, concedes that a monolithic sales force works with high-end customers who remain brand loyalists. But he believes that each of the subbusinesses will demand control of its distribution and sales in order to be truly autonomous. Stephen Smith, an analyst at Paine Webber, is direct: ''These aren't independent businesses. That's a joke. As long as a single sales force exists, it's the same old IBM structure.'' Some LOBs have already come up with ways to bypass the monolith. Cannavino claims his Personal Systems group has more sales channels than anyone else in the PC business: They include Sears, Montgomery Ward, computer stores, and its own 800 number for customers who want to order directly, as they can from numerous competitors. For the most part the M&S companies lack the kind of people they need today. Says Hixon of the Boston Consulting Group: ''IBM is being pulled in two very different directions. One is a high-standards professional services business. The other end is high-volume, high-flexibility, high-quality manufacturing with fairly low margins. But the organization in the middle -- marketing and sales -- has never been trained for either of these extreme tasks.'' At IBM North America, Bob LaBant's rather unconvincing rejoinder is that he can solve the problem through training. At least until recently, the sales force resisted selling low-margin open systems if it could push proprietary mainframes or minicomputers like the AS/ 400. Though the company won't confirm it, knowledgeable outsiders say IBM's sales force was actively discouraged from proposing the RS/6000 workstation as a replacement for aging System/36 and 38 minicomputers until mid-1991. When a computer consultant was upgrading a system for a chemical plant in Georgia two years ago, he read about IBM's new RS/6000 and decided it was exactly what he needed because it would run the same software he used on the company's old NCR machine. ''I called the local IBM office in Augusta, and they wanted to sell me an AS/400,'' says the consultant. ''They made me go around and around. It wasn't easy to convince them the RS/6000 was what I really needed.'' He persisted, saving his client about $1 million.

STRESSING SERVICES B- Services made up only 9% of IBM's revenue last year. To remain anywhere near its present size, the company will have to increase that percentage rapidly. Says Scott Flaig, a top computer industry consultant at Ernst & Young: ''Technology is becoming a commodity, and the difference between winning and losing comes in how you deliver that technology. Service will be the differentiator.'' Adds Esther Dyson, who edits the respected industry newsletter Release 1.0: ''All the advantages of being IBM -- it's ubiquitous, it's big, you can trust it, it's not going away -- don't necessarily make it better at delivering customer satisfaction, which comes from successful systems integration. As a customer, I want a supplier who's going to make all my stuff work together.'' IBM's worldwide services czar, Bernard Puckett, runs the LOB called Application Solutions, which supplies software to solve customers' specific business problems. Puckett says he's trying hard to push a services approach: ''The job is to understand the customer's needs in detail. We need people who know the customer's list of issues.'' As one important step, he helped create the first IBM consulting organization. IBM hired Robert Howe from Booz Allen Hamilton in 1991, itself a revolutionary move in IBM's solipsistic culture. Almost overnight Howe created one of the world's largest consulting firms, with 650 professionals in 40 countries, drawn from IBM's experts on manufacturing, banking, and other disciplines. IBM claims to be moving toward an industry-specific approach, and to some extent it is. Since last year, for example, every banking customer in IBM's mid-Atlantic region has dealt only with a group that specializes in banking. Previously each bank worked with the branch nearest its headquarters. And IBM recently created ''market-focused communities'' that include experts from both the sales companies and the LOBs in, say, the aerospace industry. But most of IBM's industry structure is still locked inside a geographic prison. Its banking experts in Germany have little incentive to cooperate with those in the U.S. That leaves global customers without sufficient global support. Bob Djurdjevic of Phoenix, whose respected Annex Research tracks IBM's often erratic pulse, thinks Big Blue should emulate EDS, which has reorganized since 1989 into 38 worldwide industry groups. He explains how he sees the difference: ''In IBM people are experts in product and try to learn about industries as a secondary calling. In EDS, they are primarily industry experts and their product knowledge supports that.'' CEO Les Alberthal of EDS agrees: ''The mind-set of people in a true service environment starts off with 'the customer is always right.' That's a big difference from a partial customer commitment aligned with a commitment to also move your box.'' While IBM is pushing in the right direction, Akers has barely confronted the challenge an industry-services approach represents to IBM's old ways of thinking.

EMPHASIZING CUSTOMER SATISFACTION C+ Big Blue is hardly alone here. It's the vogue of the decade, but IBM's success has been only middling. For years the company was spoiled by customers who would buy only from IBM. Some still do, since a lot of comfort remains in that Big Blue blanket. Says James Cassell, a technology analyst at the Gartner Group: ''IBM has had the Field of Dreams mentality -- 'If you build it, they will come.' '' But in the ever more real world IBM must compete in, customers are getting tougher and tougher. Says Michael Dell, founder of the mail-order PC company that is the envy of the industry -- partly for its intense customer focus: ''The people we call the 'Big Blue iceberg' who will only buy IBM are fewer and fewer every year.'' Two of the seven measures on which Akers and his corporate colleagues judge and reward the LOB managers are customer satisfaction and quality. That emphasis seems to be making some difference. Les Alberthal of EDS, who claims to be IBM's biggest commercial customer, says a top Big Blue manager, vice president David Thomas, arrived this spring for the first high-level visit in many years. When the job of IBM sales representative to MCI Communications opened up recently, IBM asked MCI senior vice president James Zucco for his recommendation, and his candidate got the job. Says Zucco: ''Three years ago they wouldn't have asked.'' But is the pace fast enough? At Metropolitan Life Insurance, which spent $54 million on IBM hardware, software, and services last year, top hardware manager Al Giordano frowns when asked if IBM is going in the right direction. ''I'm confident, but I don't see all the evidence yet,'' he says. His software counterpart, John Merrell, adds: ''It's the right direction but not fast enough.'' Nick Donofrio met with both men this year to find out what Met Life's large system needs were likely to be in the next five years and to help focus IBM's product development plans. LaBant, the man responsible for customer contact in North America, says company studies have calculated that each percentage point in improved customer satisfaction translates into $500 million more revenue over five years. At his monthly meeting with executives who report directly to him, he says, they spend two of the four hours discussing customer satisfaction and quality. He's also big on measuring quality improvements. Billing errors were reduced 33% over the two years ended in December, and the average time needed to fix software dropped from 22 days to eight. So why a C+? Customer focus will have to become even sharper if IBM is to make anything like its historic profits. Perhaps IBM ought to take a cue from rival NCR, where every corporate officer, including CEO Gil Williamson, is personally responsible for a couple of big customers.

CUTTING STAFF WITHOUT LAYOFFS A- By the end of the year Akers will have cut the work force by 85,000 from its 1986 peak of 407,000. He'll say goodbye to at least 20,000 this year alone. Managers have dropped from 50,000 to 36,000. Most departures were staff and headquarters employees. The percentage of IBMers who contribute directly to the bottom line -- those who make, sell, or service products -- has risen dramatically, from 43% in early 1986 to 57% at the end of 1991. Personnel director Walton Burdick's staff has shrunk 90%, from 400 in 1987 to 40 today. Renato Riverso brags about trimming European headquarters from 2,000 to 194.

Nobody has been laid off. One of the most cherished elements in the IBM culture is its so-called full employment, or no-layoff, policy. Instead, it has run 87 voluntary buyout and early-retirement programs in 37 countries. Few IBM practices draw more scorn on Wall Street. Daniel Benton of Goldman Sachs * scoffs: ''People at successful computer companies don't expect lifetime employment.'' Cassell of the Gartner Group says full employment means the best people leave: ''Voluntary packages mean you lose experience and those who are so good they can get jobs outside.'' In fact IBM has had a hard time holding on to some resident geniuses. The company likes to call its most creative nonconformists ''wild ducks,'' because they don't fly in conventional IBM formation. But those with the widest wingspan have frequently flown off to competitors. Andrew Heller, who ran the RS/6000 development team, once drove his vintage BMW motorcycle off an elevator and down the fifth-floor aisles of IBM sales headquarters, right past his dumbstruck boss. He left in 1989 in disgust over how long IBM was taking to bring his inventions to market -- more than ten years in the case of some devices that connect mainframes to disk drives. He now runs a Silicon Valley startup trying to develop a full range of computers to compete with IBM. Ever the bad boy, Heller named his company HAL Computer Systems. HAL, you will remember, was the murderous computer in 2001: A Space Odyssey. Akers values loyalty highly. ''I believe layoffs change the culture of a company completely,'' he says. ''The relationship between the enterprise and the individual changes forever, and boy, I don't want that change if we can possibly avoid it.'' Burdick says the company has studied the matter closely and determined that its method costs no more in the aggregate than layoffs at comparable companies. IBM now methodically ranks each employee's performance; Burdick says 70% of those who have left this year fell in the bottom half. IBM also has become much tougher on poor performers. It fires folks who don't deliver more often than it used to. In any case, says Burdick, full employment has never been written in stone. He and others clearly signal that if the Baby Blues eventually find it necessary to bend on this issue to remain competitive, corporate headquarters may be unhappy, but it won't necessarily stand in the way. Says a director, anonymously but unequivocally: ''Divisions will have the right to make decisions about layoffs.'' Work force experts are only beginning to quantify the damage that layoffs and the accompanying resentment and fear can do to an organization. Akers deserves credit for bucking his critics and trying to preserve one of IBM's distinctive qualities. At the same time, knowing they must perform to keep their jobs is a fundamental motivator for employees. Akers is finding a delicate middle ground, while making cuts that will mean huge savings over the long term.

REDUCING COSTS B+ Here there is indisputably progress, and results are quantifiable. IBM's total costs grew 12% a year in the mid-1980s. In 1991 that rate fell below 4%. If that low rate of increase continues and revenues grow more than that, profits will start climbing as well. Nonmanufacturing expenses alone -- principally sales force, rent, and administrative expenditures -- will fall 3% to 4% this year and next, according to Akers, saving $2 billion in all. The single greatest saving has been in payroll. Cost cutting is crucial, he says: ''We grew up in the past 25 years meeting me-too competitors in the marketplace and selling the value of the IBM company at a big price. If we can't, then we have resources we can't afford.'' To get rid of staff positions it sees as not part of its core operations -- benefits processing, hiring, and maintenance, for example -- IBM is setting up subsidiaries and joint ventures around the world to sell these services to itself and others. This spring a subsidiary called Workforce Solutions took on 1,400 IBM U.S. human resources people, including those who administer pay and benefits, coordinate international assignments, and run management training centers. It will sell their services to various parts of IBM in arm's-length deals and eventually look for outside clients too. In the future the LOBs will be allowed to buy these services elsewhere, if they choose. Burdick says that by consolidating all those services he eliminated 400 jobs. In time, he adds, the pay and conditions of employment for those remaining will fall to levels appropriate to the personnel services industry, not the higher levels of the computer business. IBM has historically had the Rolls-Royce of benefit plans. That's changing too, but slowly. On July 1 its U.S. medical plan stopped covering the first day of hospitalization; previously, it had paid 60% of the first-day cost. Burdick proudly describes how the company cut the annual growth of benefit costs from the mid-teens several years ago to just under 10% in 1991 -- a notably lower percentage than at many corporate giants. Managers are facing tough decisions to close facilities. Akers says IBM has trimmed 25% of its manufacturing capacity since the mid-1980s. Donofrio recently shut two of his three mainframe development labs, one in Germany and the other the Glendale laboratory in Endicott, New York, where many of the company's most successful mainframes were developed. To Akers, Glendale represents ''the old, confident, mature, family heritage of IBM.'' But he unequivocally supported Donofrio's proposal to close it. One of IBM's greatest assets continues to be its tremendously innovative research labs, by far the largest and costliest of their kind in the world. Three Nobel Prize winners walk the halls. Recently a group announced it had built the world's smallest transistor, a mere 1/75,000th of the width of a human hair and 20 times tinier than anything before it. IBM's problem has never been inventing wonderful stuff, but rather getting it out of the lab into a profitable product. Last year's $6.6 billion pricetag for R&D is simply too much. Akers has to find a way to bring that figure down drastically, even as he speeds commercial innovation. Let's be clear -- IBM was fat and, until very lately, unreasonably, ignorantly happy. It had a lot to cut. As recently as last year, temps hired in one Westchester office got only 45 minutes of work each day. When one asked to do more, she was told to bring a book like everybody else.

SHARING RISKS AND GAINING EXPERTISE THROUGH PARTNERSHIPS B+ Several years ago, IBM started linking up with competitors to offer better products and services to customers. The theory is that for many complicated tasks, no one company has the smarts or resources to go it alone. Even as Apple continues to battle IBM for PC market share, the two companies are collaborating on the next generation of chips for workstations. There are thousands of less visible ties. IBM North America has over 4,000 partners, many crafting highly specific software for smaller customers. Some matchups should save IBM tremendous time and expense, and possibly help create vast new businesses. In semiconductors, the company has joint development projects with the country's three biggest chipmakers -- Motorola, Intel, and Texas Instruments. Michael Attardo, general manager of the Technology Products LOB, says these ventures guarantee IBM access to microprocessor innovations. ''By 1995,'' he predicts, ''we'll be able to create a workstation on a chip.'' Attardo adds that partners will be essential in developing denser and denser memory chips as well as microprocessors. For the next generation 64-megabit chips, Attardo says a single plant will cost about $1 billion, which no company can afford alone. The generation beyond that -- 256 megabits, 64 times denser than today's standard high-end PC memory chips -- may mean moving beyond current technology to X-ray lithography. IBM owns the only commercial X-ray synchrotron in the U.S. Says Attardo: ''Any number of our Japanese friends are interested in working with us on X-ray lithography. You will see us working with two or three other companies. We may announce something within six months.'' And by the way, who says you can't marry your partner? Akers says he can envision buying a big company in the not-distant future, something that has been foreign to IBM. ''As the industry evolves, why not?'' he asks. ''If you want to impact a $60 billion revenue stream, you've got to do something beyond investing a couple hundred million dollars here and there.'' OVERALL, a grade of B-. The company's efforts are slowly gaining fans. Says Philippe Kahn, the outspoken president of Borland International, which works closely with IBM on PC software: ''There's been a profound change at IBM in the past year. They're becoming more competitive, more cutting-edge. It's surprising they've had the guts to make some tough decisions.'' Silicon Valley marketing consultant Regis McKenna, whose book Who's Afraid of Big Blue? appeared in 1988, says he too is more optimistic today: ''I think IBM is turning itself inside out.'' Analyst Barry Bosak of Smith Barney expects ''explosive'' profit growth and an upgraded dividend as early as next year as a result of Akers's cost-cutting efforts. Board member Stephen Bechtel Jr. bought 5,000 shares in May. Akers himself won't get his final grades for several years. Unlike many other computer industry CEOs, the man is no techno-visionary. After all, he rose through IBM's ranks as a salesman. This spring rumors abounded that the board was about to replace him with Jack Welch of GE -- a change some in the industry would welcome. Says Raymond J. Noorda, chairman of Novell, the leader in networking software, and an ex-GE man: ''They need a Jack Welch mentality for a while.'' Earlier reports had Ross Perot stepping into Akers's shoes.

Akers is passionate about trying new ways of doing things and leaving subordinates, some of whom do have that visionary streak, to manage without his intrusions. He is showing other intriguing signs of boldness. John Kotter, a Harvard business school professor, has studied cultural change in large ! companies and concluded that corporate insiders like Akers can seldom transform an organization beset by inertia. To Akers's credit, he invited Kotter to discuss his research with top IBM executives at the annual corporate strategy session in May. ''I came out of there a little more optimistic than I went in,'' Kotter says. ''I confronted them with some really tough stuff, which could get them defensive. But people started stepping up to the plate and saying, 'This is us, guys.' Akers ended by saying, 'Take us one more step.' I said if I had a magic wand I'd just crank up the urgency dial inside all of you. He came up to me afterward and said, 'You've done exactly what I wanted today.' '' Akers argues that IBM is so complicated that an outsider would take too long to learn its ins and outs. As for himself and his team, he says, ''At least consciously, we do not feel inhibited at all about change. We're prepared to change anything except the values of excellence, of customer service, and of our respect for one another.'' The key question: How far will Akers crank up the urgency dial?