What Lou Gerstner Could Teach Bill Clinton.

Lessons for government from IBM's dramatic turnaround

IF STEPHEN KING WERE TO WRITE A NOVEL about the federal government, he could scarcely do any better than the tome that landed on President Clinton's desk on June 15. The Report of the President's Foreign Intelligence Advisory Board was destined to be grim reading: Its authors were asked to look into security lapses at the Energy Department following last winter's revelation that China had stolen high-level nuclear secrets. But what former Senator Warren Rudman and his fellow panelists saw was clearly beyond their worst fears. "Never have the members of the Special Investigative Panel witnessed a bureaucratic culture so thoroughly saturated with cynicism and disregard for authority," they write. The DOE hadn't simply let a few clever Chinese spies make off with classified papers. It was virtually inviting people to pilfer the material and know-how to make nuclear weapons. Although its scientists are among the best in the world, the Department's bureaucrats deliberately "defeat security reform initiatives by waiting them out." A honeycomb of bureaucracy has "diffused responsibility to the point where scores claim it, no one has enough to make a difference, and all fight for more." Furthermore, these problems "have been cited for immediate attention and resolution ... over and over ... ad nauseam." The authors conclude that the Department is simply incapable of maintaining nuclear security, and they propose that its weapons research and stockpile management functions be turned over to a new semi-autonomous sub-agency within the Department, or to a new independent agency.

You might suppose that any self-respecting corporate executive, viewing this catastrophe from the comfort of the breakfast table, would declare that only government could screw up so badly. Yet at least one former high-level IBM official saw a certain parallel between his former employer and the DOE. After all, it was only six years ago that IBM, once the world's premier computer company, was teetering on the brink of collapse. Between 1991 and 1993 the company had lost $15.4 billion, and sales of mainframes, its old standby, had dropped by half. These failures were not simply the reflection of a changing marketplace. Like the Energy Department, Big Blue was burdened with a vast, redundant bureaucracy that was virtually guaranteed lifetime employment and resisted change with all its might. It too had a culture of arrogance rooted in its record of scientific superiority. And like Energy, IBM had virtually given away some of its best technical assets to its competition--in this case Microsoft.

Finally, IBM too had been diagnosed ad nauseam, and many demands for reform had come and gone. Many computer industry observers figured it was curtains for Big Blue, and arrangements were being made to subdivide it (again, like the Energy Department). Even Lou Gerstner, the CEO who later came on board and ultimately turned the company around, was close to despair. "It just looked like it was going into a death spiral," he said later. "I wasn't convinced it was solvable."

Six years later, IBM is utterly transformed. The civil wars that crippled the company are over, and the arrogance that blinded it is gone. The covers of IBM's annual reports accurately reflect its internal change: Blue-suited men walking through a maze of computers have been replaced by smart-looking young women in bluejeans. Most importantly, the changes have paid off. In 1998 IBM had record revenues of $81.7 billion, and a profit of $6.3 billion. Its market capitalization has grown roughly tenfold since Gerstner took over, and Wall Street analysts expect the company's stock (and profits) to grow by at least 20 percent over the next year.

Time to start running the government like a business? Well, no. We've all grown tired of CEOs who think they can hop over for a stint in government and "kill that snake," as Ross Perot used to say. But the IBM story isn't your typical corporate fable. Lou Gerstner didn't just cut costs to make shareholders smile, or mouth the latest corporate slogans. His successful turnaround has already become the stuff of legend. Yet no one has bothered to point out that the parallels between Big Blue's problems and those of the federal government are uncanny. The excitement IBM's revival has stirred says a lot about the passion we all seem to feel for business stories at the end of the millennium. It also says a lot about what government might achieve if we could bring the same passion to its renewal.

Big Brothers

It's no accident that IBM and the federal government have had similar illnesses, because they grew up together. Thomas J. Watson, a former traveling salesman of organs and sewing machines, started his computing career with a punchcard accounting system that had been developed for the 1890 census. He changed his company's name to International Business Machines in 1924, and as the federal bureaucracy grew during the Depression, IBM grew with it. The Social Security Act of 1935 and the Wages-Hours Act of 1937 required companies to log the hours worked, wages paid, and overtime earned for America's 26 million workers. This was big business for a maker of tabulating machines and time clocks. IBM also mimicked the government's paternalistic solicitude: The company was among the first to provide group life insurance (1934), survivor benefits (1935), and paid vacations (1936), writes Robert Slater in Saving Big Blue. After the war, IBM's virtual lock on the nascent technology of computing made the company fantastically profitable.

At the same time, IBM began to develop the distinctive culture that would lead to its rise and ultimate fall. Tom Watson and his son, Tom Jr., were perfectionists who saw the family business as a sacred trust and ran it like benevolent dictators. Employees were expected to devote themselves utterly to the company--and to its founder. There were company songs, with lyrics like this: Our voices swell in admiration, of T.J. Watson proudly sing; He'll ever be our inspiration, to him our voices loudly ring.

In return for such devotion, employees were given all kinds of perks. "It was the kind of place where they might call you in on Christmas day," recalls one former IBMer, "but then they'd send you on a paid vacation somewhere, or if your wife was sick they'd send flowers ... they took care of you" Once you joined the company it was expected that you'd be there for life--unless you violated its rules of behavior, which were enforced as rigorously as its code of excellence. IBM men didn't just wear the company uniform of dark suits, dark ties, starched collars and fedoras. They were to be moral exemplars too. "In the old days, you could be fired for serving liquor to another IBMer in your own home," recalls one employee. Divorce was frowned upon.

As the company's profits and prestige shot up through the 1950s and '60s, so did its bureaucracy. The turning point, according to Robert Slater, came in the mid-'60s, with the development of the System 360. To manufacture these computers, the company had to build five new factories, and to man them they hired armies of new people. IBM's population rose from 87,000 in 1963 to 198,000 just three years later. The company continued to grow for the next two decades, though not as quickly, reaching an all-time high of 407,000 in 1986.

By that time it had become clear to outsiders that IBM was an elephant in an industry that required a hummingbird's agility. But IBM's arrogance, built up over decades, prevented it from moving beyond the mainframe computers that had been its mainstay for so long. If IBM had seen the potential of personal computer software, it could have bought Bill Gates out when he was still a pimply post-adolescent. Instead, it gave him a virtual monopoly on its operating system and helped to create a billionaire. Big Blue might have made a similar profit off relational databases, but it waved them away, and Larry Ellison used them to build Oracle into another West Coast powerhouse. Compounding this failure of vision was a perfectionism that prevented IBM from getting its own products to market before it was too late. Early in his tenure as CEO, Lou Gerstner told his research people, "You don't launch products here. They escape."

In the late '80s CEO John Akers tried to slim the company down with voluntary buyouts (giving employees a financial incentive to leave early). But in a company renowned for lifetime employment, this helped to create a climate of fear that only worsened IBM's rigidity. "You didn't innovate," one former exec told Robert Slater. "You didn't want to do anything that would give you a high profile...."

By early 1993 Big Blue appeared to be beyond hope. It had taken more than $20 billion in losses, and its debt rating was dropping fast. The board of directors sacked John Akers as CEO and dangled the job before several corporate saviors, including GE's Jack Welch, AlliedSignal's Lawrence Bossidy, and Eastman Kodak's George Fisher (then at Motorola). There was even talk of asking Bill Gates to take over. No one was willing. Lou Gerstner, then CEO of RJR Nabisco, initially refused the offer. He accepted only after IBM board member James Burke appealed to his sense of civic duty: IBM was vital to the American economy, and it must not be allowed to die. It was a good line, but the smart money said that no one could bring Big Blue back to life. Gerstner reported for duty at company headquarters in Armonk, NY on April Fools' day, 1993.

No More Yes-Men

Gerstner's first task was to change the ingrained attitudes that had prevented IBM from seeing the world around it. These are often described as a product of the company's unique place in American business--as one old company saying went, "there's the right way, the wrong way, and the IBM way." But IBM's problems were really not so different from those of any large bureaucracy. "If you leave institutions in place for too long, whether governments or corporations, they get focused on maintaining themselves as institutions," says Jim McGroddy, who ran IBM's research labs from 1989 to 1995. "What they achieve for the customer becomes very secondary."

At IBM, as in the federal government, this meant that people who made criticisms--however valid--were often ignored or even punished. "The operating principle was, don't make the boss unhappy," says one longtime IBMer. This attitude, she adds, did incalculable damage to the company. Legitimate criticism of products and strategy never percolated up to the people who needed to hear it. And managers would take a financial hit rather than admit a mistake. In IBM Redux, Doug Garr describes an IBM supplier who tried to return a $20,000 overpayment. IBM refused to accept it, because in order to do so someone would have to admit an error was made.

Gerstner witnessed this "good news only" syndrome at one of his early meetings, where he heard a presentation from the chip-making division, which was widely known to be having problems. The head of the chip division told Gerstner that the problem was that there was no mainframe business, so his division couldn't make money making chips for mainframes. Gerstner stopped him right there and asked why the mainframe guy had just told him his division was doing fine. As it turned out, the mainframe division was in trouble--but they couldn't bear to admit it to the chief.

"Gerstner made it very clear very quick that you tell the boss the truth," says one IBMer who witnessed the turnaround. How? For one, he spent much of his time during the first few months talking to IBMers, consultants, and customers about the company and making it very clear that he wanted unvarnished truth. "He was the best listener I have ever seen," says Sam Albert, an IBM analyst who worked for the company from 1959 to 1989. Early in his tenure, Gerstner sat down with Albert, and as he asked him what he thought about the company, he took five pages of handwritten notes--not exactly typical CEO behavior.

Perhaps Gerstner's most effective tool for enforcing honesty in the company was the way he conducted meetings. In the old IBM, meetings were "like a high mass, with supporting documents as thick as your arm," says Dan Mandresh, who was Merrill Lynch's chief IBM analyst for 20 years. The most famous ritual at these meetings were the visual-display "foils" execs used to project charts and graphs onto the wall. "We used to describe talks as a `six-foil talk' or a `four-foil talk,'" one veteran recalls. "There were IBMers who literally didn't know how to talk without foils." Meetings were too large, because people were included based on their rank, whether or not they had anything of substance to add.

In the government it's the same story. For many federal bureaucrats, meetings are an end in themselves, a way to create the impression of activity (and taxpayer money well spent) where there is none. Being invited to them is a reflection of rank and status, so that people attend them regardless of whether they have anything to contribute or learn. At the Energy Department, for instance, meetings of the Office of Strategic Planning, Budget, and Program Evaluation have been attended by dozens of representatives from other offices, whose job is often confined to writing notes about the meeting for their own offices--regardless of whether it has any relevance to their own work.

Gerstner recognized that all this pointless ceremony was crippling IBM. One of the first things he did was to convene his top 20 executives and tell them all to write a short paper, with no visuals, answering these questions: What is your business? Who are your customers? What is your marketplace? What are your strengths and weaknesses? Who are your main competitors? He told them to get it done in two weeks, and that he would meet with each of them one-on-one shortly afterward to discuss it.

That may sound pretty reasonable to an outsider. At IBM, it was revolutionary. "People weren't used to writing in sentences," recalls Jim McGroddy. When he met with Gerstner, "it was just the two of us standing at a table. No projector." The meeting went well; McGroddy convinced the new boss that he was willing to help the company change.

Others, however, tried to bring out the foils, and "Gerstner jumped all over them," says McGroddy. According to IBM lore, Gerstner actually walked up to the projector at one meeting, turned it off, and told the exec, "If you can't explain it to me in your own words, you don't understand it." Before long the foils were gone. With them went all the old rituals that had made meetings such a waste of time. He insisted that meetings be as small and as quick as possible. To many company veterans, this felt like a slap in the face. But Gerstner made his intention clear: "I am trying to avoid the concept of reward being associated with these meetings."

Firing at Will

The next serious challenge Gerstner faced was reforming the company's workforce. "Unless you did something really bad, they didn't fire you," says Sam Albert. "They'd just move you to another job." This too is an all-too-familiar tradition in the civil service and in public schools, where administrators "pass the lemons" to another division or school district because they cannot fire them without going through a near-interminable series of hearings and appeals.

John Akers had downsized the company significantly starting in the late '80s, but he had done so primarily with generous early retirement incentive plans. The trouble with this method was, as one IBMer put it, that "some of the best people were tempted to take it--because they were the ones who knew they could get work elsewhere. The less talented people were more likely to play it safe and stay." The government suffers from the same problem. Al Gore and his "reinventing government" staff often boast about how much they've cut down the rolls. But they've used the same methods, with the same ill effects: The people who should be fired stay on, while those who do real work leave.

How did Gerstner handle this? IBMers claim he put the fear of God into anyone who might have been slacking off. "We had a few public hangings of people who didn't want to get on the new programs," as Gerstner later put it. "That told everybody we were Serious." Most of the layoffs that occurred after Gerstner arrived in '93 were achieved by lopping off whole units that were deemed unnecessary. But in order to make sure that IBM downsized effectively, Gerstner also instituted a new evaluation scheme. The old system, like everything else at IBM, was collegial: a friendly annual report by your manager with grades--on a scale of one through four--that were by all accounts highly inflated. Gerstner added a new element, known as "360 degree feedback": a half dozen peers of your choice would fill out confidential evaluations. Also, the annual performance grades were put on a curve within the division. Many employees complained that this system was unfair; one division might be all stars, whereas others deserved to be junked entirely. Still, the new system gave managers the tools to fire poor performers, and it kept everyone else on their toes. "The corporate entitlement culture was gone," says Ken Thornton, general manager of IBM's public sector department. "The slogan became `happy to be here, but prepared to leave.'"

Gerstner also recruited far more aggressively than IBM had in the past. During its heyday, "We just used to wait for the resumes to come in," recalls one longtime IBMer. "Now the competition for good people is extreme." IBM's stodgy image didn't help: By the early '90s ambitious young people with computer skills were far more likely to head for hipper, younger companies on the West Coast. There is a parallel with the government here too, though the federal recruiting problem is more long-standing and infinitely worse. The quality of its new hires has been plummeting for decades, and a 1998 study by the Merit Systems Protection Board cited federal superintendents Who claimed that "their assessment of applicant quality had fallen for just about every type of job category." Amazingly, the government has done virtually nothing to reverse this trend. Few agencies bother to recruit on college campuses, and some don't even answer phone calls from prospective applicants.

IBM was never in such dire straits, but they've made a massive effort to reach talented young people. A series of stylish ad campaigns harping on IBM's leadership in Internet technology gave new shine to the company's image. They also bit the bullet and began recruiting as aggressively as possible: The company that once sat back and waited for the world to come to it now sets up tables at spring break in Florida.

The Bigger the Better

Perhaps the greatest challenge Gerstner faced in his first year was finding a way to make the company work as a single team again. Most people believed it was impossible. When Gerstner arrived, the industry was near-unanimous in its belief that IBM was too big to survive, and that its components would be leaner and more effective by themselves. In fact, John Akers had already begun to implement the plan, dividing IBM into 14 separate companies--the "Baby Blues," echoing the breakup of Ma Bell into Baby Bells in 1982. Again, the parallel with the federal government was clear. Faced with a stubborn bureaucracy, reformers tend to go for the obvious solution: chop it up, or create separate subagencies to reallocate responsibility.

Gerstner was not convinced, so he convened a conference of IBM's top 200 corporate customers in Chantilly, Virginia, and asked them point-blank: What kind of IBM do you want? The answer was clear. They wanted one-stop shopping, someone who could offer hardware, software, and services. No one else could do that. "Look at what many of our competitors are doing--buying each other, striking deals and alliance so they can offer more pieces of the solution," Gerstner told Software magazine in 1997. "In many ways, they're trying to cobble together a lot of what IBM already has."

Having made that decision, Gerstner was faced with the fearful task of making IBM's units work harmoniously. "One of the key things in the federal government is the ability to scale across heavily entrenched silos," says IBM's Ken Thornton, who manages IBM's business with the government. The challenges at IBM were very similar. Under the old system, there was no incentive for IBMers to think beyond their business unit. Naturally, this exacerbated the tendency of the divisions to wall themselves off and develop products without regard to the company's larger goals. Sometimes IBM units actually resisted directives from above, much as the Energy Department's bureaucrats do, in what was known in the company as "pushback." Often this led to an idea getting stalled in the bowels of the company for so long that it never saw daylight.

Exacerbating these failures were IBM's elaborate protocols for transactions between its units--so elaborate that the company required some employees to take a two-day training class in procurement (until recently, the federal government did the same thing). "Now wait a minute, we all work for the same fucking company," said Jerry York, Gerstner's aggressive new CFO, when he heard about the problem.

Gerstner started off by pegging 40 percent of every employee's bonus pay to the performance of the overall company, as opposed to the business unit. Individual bonuses were also tied to performance evaluation, which was determined in part by something called "team," meaning "how well you leveraged IBM's overall resources in your work," says IBM's Jana Weatherbee. "But it wasn't just the money," says one former IBMer. "He made heroes of people who did it right," with reward ceremonies and prizes. Gerstner introduced a new mantra at the company: the customer first, IBM second, your business unit third. And he reinforced it by dragging the scientists who dream up new products out of the lab and having them meet with customers.

This was a radical change for IBM and long overdue. In the old days, the labcoats were so out of touch with the marketing division that their products would often emerge too late, or in unmarketable form. "I remember being at one meeting, we were at a swimming pool, with cocktails, in the evening," recalls Jim McGroddy. "There was this guy from a bank, a customer. He picked up a radio-frequency LAN adapter (which IBM had on display) and he said, `This is a really neat piece of technology. It's completely useless to me. Let me tell you what our problem is.'" He proceeded to describe the device he and bankers like him needed and said he would buy thousands if IBM would only make one.

To get the divisions talking to each other again, Gerstner had to do more than force them to start talking to end-customers (though he did do that, and it has made an enormous difference). He had to find a vision that would inspire and unify the whole company--something federal agencies also desperately need. Although he had initially (and famously) poopooed the need for a vision, in 1995 he began to position IBM to lead the next wave of "network-centric" or "pervasive" computing. An internal IBM document recalls that the industry "laughed" at Gerstner's claim that in the future "the Internet will be about business, not browsing--and working, not surfing"

Four years later, Gerstner's vision appears to be materializing. IBM has taken the lead in a new generation of "smart devices" embedded with powerful microprocessors. These will allow refrigerators and vacuum cleaners to tell their manufacturer when a part is about to fail and order a replacement. Already, some city agencies are using a device that allows them to contact police, fire, and emergency personnel faster than ever before through connected pagers, cell phones and wireless "personal digital assistants" It may sound like Star Trek, but they're not laughing on Wall Street, where IBM's stock price is surging.

Take a. Memo

What lessons does IBM's Lazarus-like return from the dead offer for government? There are several, and implementing any one of them would go a long way to restore the dignity of public Service.

The Truth Factor. Just as Lou Gerstner cracked IBM's "good news only" culture, the feds need to ensure that employees are rewarded, not punished, for sticking their heads up and delivering honest criticisms of their organizations. By the same token, appointees and managers will have to start reaching down to get the unvarnished truth about what's going on at every level, instead of just hoping nothing explodes on their watch.

Meaningless Meetings. It wasn't easy to hack into IBM's slow-motion meetings and everything that went with them: "too many sign-offs, too many reviews, too many task forces, which add work but not much value," as Gerstner once put it. But he did it. The situation is even worse in federal agencies, but there's no reason it can't be reversed there too.

No More Sectarianism. Federal bureaucrats are encouraged to look out for their agency or even their unit or office, often regardless of whether those goals have anything to do with the larger public interest. They're ripe, in other words, for the kind of transformation Gerstner pulled off at IBM when he forced his people to overcome their notorious internal divisions and work for the good of the company.

Measuring Performance. Gerstner looked to other large companies to find an evaluation scheme that would keep people on their toes and help to identify those who should be fired. The government needs to do the same. There have been some efforts on this front; the Clinton administration's latest was just six months ago. But all have fizzled in the absence of strong political will.

Hiring the Best. IBM recognized quickly that it couldn't get back in the game without drawing on the best people at every rank. The government has been abysmally slow to see this logic, but there's little doubt that few of the goals listed above can be achieved unless the feds take strong measures to reverse what one political scientist has called a "death spiral" in the quality of the government's new hires.

None of these things will be easy to achieve. Unlike IBM, the government is never far from a 535-headed monster called Congress, which often encourages reform with one hand while stifling it with another. The civil service rules themselves are a considerable obstacle, as are the federal employee unions. Real reform will also require strong direction from the president. Al Gore has shown considerable interest in this task, but with the exception of some dramatic downsizing, his reinventing government initiative has thus far stopped short of all the truly difficult hurdles.

Still, if you doubt that government can learn from the IBM example, consider what has happened in NASA's Jet Propulsion Lab in the past decade. JPL is responsible for designing the robots that go on interplanetary missions, such as the 1997 Pathfinder mission to Mars. Less than 10 years ago, it had a virtually unlimited budget and was doing multibillion-dollar missions, each of which took seven to ten years to go from idea to launch. When the Cold War ended, Congress' priorities changed, and suddenly JPL was being asked to design smaller robots, a lot faster, and with a much more limited budget.

Like IBM, JPL used to work in divisions, handling the various different parts of the robot: communications, power, structures. They'd do their budgeting and design separately, then convene, and if there were mismatches, they'd go back to their separate drawing boards, and convene again as many times as needed. It was a costly and time-consuming process. That was OK in the old days, when size and money didn't matter.

But NASA administrator Dan Goldin had imposed strict cost constraints, creating the kind of pressure private companies like IBM face. "If you're 15 percent or more over budget, your existence is called into question," says Ivan Rosenberg, a consultant who works with JPL. The engineers began to work together, and their efficiency improved exponentially. Writing proposals for new projects, which used to take them months, now takes two to three weeks, and more of them are successful. Like IBM, JPL transformed the way it conducts senior management meetings, and what used to be an all-day meeting now takes two hours. These changes have had an enormous effect on the big picture: Since 1992, JPL's budget has risen by only about 5 percent, while its workforce has been cut from 7,500 to 5,000. At the same time, it has gone from one launch every five or ten years to almost a launch a month.

Other federal workers may not be able to change as quickly as those at JPL, who are literally rocket scientists. Nonetheless, JPL's achievement is proof that stagnant bureaucracies can be reformed, even in the public sector. That's why the current cynicism about government reform is so discouraging. Rudman and his fellow panelists are right that the Energy Department is a mess. But they're wrong to give up on changing it. Their proposal to hand over nuclear security to a new subagency, which the Senate is now considering, is a gesture of despair. It would leave the rest of the Department untouched, and the new agency would surely complicate DOE's residual task of cleaning up nuclear waste at weapons sites. "We run the risk of spreading the virus we're trying to kill," says Don Kettl, a University of Wisconsin political scientist who has spent years studying the DOE.

That would be more than just a shame. Attacking Energy's management problems may not be as dramatic as a corporate makeover. No one will become a millionaire in the process, and there won't be neat gadgets in it for the rest of us. On the other hand, if nuclear waste starts to bubble up in your backyard, or a terrorist detonates a neutron bomb on Wall Street, we could all end up wishing we'd shown more interest in what those government bureaucrats do all day.