Editor's note: Every Sunday Fortune publishes a favorite story from its magazine archives. This week we turn to a 2004 Fortune profile of Microsoft CEO Steve Ballmer, who will retire within the next 12 months after more than 30 years at the company.
Photos by Antonin Kratochvil
How do you impose order on a giant, runaway Mensa meeting? Just watch Microsoft's CEO.
Steve Ballmer is a man whose reputation exceeds him. Barrel- chested and bombastic, he's always been the quintessential, larger-than-life, rah-rah leader, and the perfect foil for his geeky and erudite best buddy, Bill Gates. Stories abound of his over-the-top antics rallying Microsoft's troops at corporate events: There was the infamous Monkey-Boy dance he performed at a sales meeting a few years ago, which prompted dozens of incredulous Microsoft bashers to post choppy videoclips of his cavortings on the web; or the surgery he endured to repair blown-out vocal cords resulting from a bout of unrestrained cheerleading; or the outlandish hula-girl and used-car-salesman costumes he gleefully donned to get a laugh and make a point. On the Microsoft campus, you can always tell he's approaching by his sharp, punctuating claps as he barks orders to a subordinate trailing behind. And he's the first to admit that in his case, "table pounder" is literally apt.
That noisy behavior served Ballmer--and Microsoft--quite well over more than two decades as Bill's go-to guy, whether it was as leader of the Windows software development team in the 1980s, as chief of worldwide sales and marketing in the 1990s, or in his stint as president. Ballmer had a big hand in shaping Microsoft's take-no-prisoners corporate culture during its meteoric rise, which made him a billionaire several times over. The surprise is what he's done in the four years since Gates made him CEO. You'd have expected more rabble-rousing, and there's been plenty of that. Yet at the same time Ballmer has steered Microsoft through the tech bust (Gates, with impeccable timing, made the handoff just three months before the dot-com bubble burst). Even as the rest of IT shrank, Microsoft continued to grow and to maintain prodigious operating profits of more than 35%.
More important, Ballmer has put his stamp on the company. After an admittedly tentative first year as CEO, he embarked on a methodical campaign to rewrite Microsoft's corporate "operating system" from top to bottom. He's taken an unwieldy, overcentralized company that flew by the seat of Bill Gates' pants, and broken it into seven operating divisions to push decision-making and accountability out into the organization--each unit has its own P&L. At the same time, he has personally overseen the creation of a playbook of procedures to systematize everything Microsoft does, from product development to strategic planning to employee and management evaluation, as well as softer things like the company's mission and values. He has completely revamped the compensation plan, and last year eliminated stock options altogether in a move that stunned the tech world. He has ordered up crash projects to tackle Internet security and spam. It's all a work in progress, but two years after the overhaul began in earnest, Microsoft is beginning to behave like a grownup company instead of a raucous, runaway Mensa meeting.
And if there's one thing Microsoft needs now, it's steely discipline. In recent months revenue growth has slowed to single digits; despite the Nasdaq's 49% run-up in 2003, Microsoft stock stayed flat. The company is facing its biggest competitive challenge ever. The Linux operating system and other so-called open-source programs--Ballmer prefers to call them "free software"--threaten not only to make inroads into Microsoft's dominance but also to drive down the prices of all software. IBM, meanwhile, is doing everything it can to advance the perception that world-class information systems require lots of care and feeding, in the form of software integration and customization, consulting, and outsourced services. As the chief proponent both of Linux and of the value of IT services, Big Blue would like nothing better than to change the rules of the game and relegate Microsoft to being just another software company.
In the face of this onslaught, Microsoft's ambitions for what it can do with software have never been grander or more complex. It wants to rewrite the rules too--by increasing the perceived value of software. It knows the only way to do that is to build software that works so well "out of the box," that is so well integrated from the get-go and so rich with useful features, that it renders IBM's armies of consultants extraneous and makes Linux and other "free software" look banal.
Bill Gates, who is now ensconced as chairman and chief software architect (and who owns of 5% of Microsoft stock), is determined to make Microsoft code even more ubiquitous, linking all kinds of computing and communications and consumer electronics devices in a mesh of software that will make the entire Internet and everything on it a single, programmable entity. For starters, that means overhauling Windows for servers and PCs--Microsoft's so-called Longhorn project, which Gates is leading. It also means finding ways to implant Microsoft code in cellphones, game machines, and other consumer devices--pretty much anything with an electrical current--so they will play together with a minimum of human tinkering. "Nobody has ever built anything as big as what we're trying to build now," says co--chief technology officer David Vaskevitch. "It's a humbling prospect."
Which brings us to Steve. To put it simply, his job is to refashion Microsoft into a rocket that can reach that moon. Gates hopes Ballmer's rabid team spirit combined with his fixation on imposing order will enable the 55,000-person company to "scale up." Says Ballmer: "We need a new framework to make what is now a very large organization greater than the sum of its parts." Echoing the thought, Gates says the organizational challenge is "getting the IQs to add up, and having no subtraction." He adds wryly, "You know, we're paying for all that IQ." Listening to that remark, it's not hard to imagine that the heat on Ballmer isn't ust from free software and IBM.
You have to hand it to Gates: he did a pretty good job of growing Microsoft from a two-man partnership into the world's most valuable corporation at the point when he relinquished the CEO job in January 2000. (The market cap then was some $600 billion, vs. $304 billion now.) A notorious penny-pincher and control freak, Gates was also an instinctive entrepreneur and company builder. "Bill masterminded all of this--from an unmanaged, wild startup to a company that never had a bad quarter in 28 years," says Vaskevitch. So it's not surprising that Gates is a little defensive about the state of the company when he stepped down: "It's not as if when Steve took over there were receipts all over the place and everything was out of control."
Ask senior Microsoft lieutenants how they felt, though, and some paint a bleaker picture. "It was a much tougher period than most people realize," says CFO John Connors--the Internet craze, by causing business to boom, helped hide the problems. Windows development chief Jim Allchin is blunter: "It was very frustrating. There were a few of us saying, 'It's not working, it's not working, it's not working.' Not only people like me, but on down, because we just couldn't scale" to meet the demands of crucial software development projects. Adds group vice president Jeff Raikes: "You can't expect one guy or two guys to sit on top of $5 billion to $7 billion of R&D and make decisions on a project-by-project basis."
The first year under Ballmer wasn't pretty. Microsoft did fine financially--sales and profits were up 10% and 6%--but its people were chastened by the company's legal battle with the Justice Department and shaken by the IT slump and by a creeping fear that Microsoft might turn into just another megacorporation. Scores of experienced managers and engineers in their 30s and 40s--many of whom were financially set for life, thanks to having cashed in stock options during the boom--made for the exits.
Ballmer's first big initiative as CEO--a product and marketing blitz known as .NET--was emblematic of how sand was working its way into Microsoft's gears. Genuinely ambitious yet hastily conceived, .NET was launched in June 2000 in an effort by Microsoft to gain the upper hand in using the Internet to deploy a new kind of application called web services. The technology vision was pure Gates--it involved rejiggering virtually all of Microsoft's software. But the marketing plan, primarily Ballmer's responsibility, baffled both insiders and customers. Nobody knew what .NET really was; many thought it just a ploy to foist on users yet another software upgrade. The situation was painfully embarrassing, says Allchin: "The sad part is, there was so much goodness in the technology, but it really got quite confused. Some people in the company just started grabbing the name and applying it to existing products because they wanted to be with the program."
Ballmer, meanwhile, was finding Microsoft almost impossible to manage. Decision-making was still as centralized as it had been under Gates--Ballmer was once asked to approve the hiring of an administrative assistant to a country manager in Eastern Europe. He began to envy GE, where an old acquaintance named Jeff Immelt was soon to become CEO. As every MBA student knows, that legendary conglomerate frees its CEO to concentrate on high-impact decisions by running its businesses as autonomous units.
Worse, though Gates had made Ballmer CEO, it quickly became clear that Gates hadn't really ceded the power. For months he second-guessed Ballmer's every move. When Ballmer first proposed to segment Microsoft into divisions, Gates fought the idea, arguing that the company had to stay monolithic because there was so much interdependency between, say, the Windows PC group, the server group, and the Office group; their software had to be on the same page. "Bill and I kind of had a brouhaha because I didn't really get what was in his mind, and [the business] was so hard to manage," recalls Ballmer. "I was hell-bent to try to dial up the autonomy. It's not that Bill is anti-autonomy, but he kept saying 'This has to work with this, and it's got to work with this.'"
Inevitably, the tension between the two boiled over in meetings. Recalls Allchin: "There were some hard meetings when a few of us told them, 'You guys need to get your act together.'" Marketing vice president Mich Mathews likened the situation to a marriage on the rocks. "I felt like, 'Mom and Dad never argue like this.'"
The ice broke in early 2001, according to Ballmer. "I finally told him, 'I sort of get it.' We couldn't just have these groups siloed like GE. So we'd have to come up with a structure unlike anything out there" to simultaneously give divisions enough autonomy to manage themselves, yet make it easier for them to cooperate and integrate the technology. That concession by Ballmer reassured Gates. The second-guessing ended, and the Ballmer era at Microsoft finally began.
Boilerplate Steve Ballmer biographies never let you see what makes him so uncannily suited to organizing a 55,000-person business. The son of a midlevel manager at Ford Motor, he grew up in suburban Detroit, winning scholarships to attend a prestigious private high school and then Harvard. There he had the good fortune to live on the same dorm floor as Gates. Steve always liked running things: In school he managed athletic teams and student newspapers. After Harvard he spent a year marketing cake mixes at Procter & Gamble, where he shared a cubicle with another business neophyte, named Jeff Immelt, and then enrolled in the MBA program at Stanford University. Within a year, he dropped out at Gates' behest to become the "assistant to the president" at Gates' startup. Ballmer's greatest gift seemed to be his motivational skills, which he employed to more and more effect as he ran major development teams and later the worldwide salesforce.
Yet underneath it all--here's where the potted biographies leave off--Steve Ballmer is a management wonk. The degree he earned at Harvard is in applied math; his favorite college course was one that analyzed management techniques used by orchestras, dance companies, publishers, and the like--creative organizations in which success and productivity are difficult to quantify. To this day, the 47-year-old Ballmer is obsessed with the intricacies of business processes and decision theory, with organizational dynamics, with devising ways to measure employee and corporate performance, and with building elaborate employee-and customer-feedback mechanisms, all to get a better handle on how a large enterprise actually operates.
His approach to organizing and managing Microsoft (msft) amounted to the most far-reaching overhaul the company has ever had. Creating semiautonomous profit centers was only the obvious part; Ballmer also wanted to revitalize the culture and make explicit to his newly empowered managers exactly what they were supposed to do and how they were supposed to do it. He went about all that in trademark Microsoft style--geeky earnestness. He ordered up a new mission statement: To enable people and businesses throughout the world to realize their full potential. (The old mission statement, which Gates wrote in 1978, seemed mostly fulfilled: A computer on every desk and in every home.) He also came up with a set of corporate values he calls the "tenets" that address both personal behavior (employees must have integrity and honesty; be open and respectful to others; and so forth) and corporate goals (broad customer connection, trustworthy computing, etc.). Ballmer knows how corny all that sounds. ("I'm not nuts," he says.) But he also knows that Microsofties, for all their smarts, have always seen their work as a crusade, and that the bromides help give the endeavor moral heft.
In 2001, Ballmer and his lieutenants tackled the huge task of divvying up more than 30,000 developers, marketers, salespeople, and engineers around the world among the $9.2-billion-a-year Information Worker division; the $2.7-billion-a-year Home & Entertainment division; and so on. An immediate effect of breaking out financial figures for these units was shock. The numbers laid bare to the world money-losing operations like the MSN online service, the Xbox game machine, and software for cellphones and PDAs. Too bad, says Craig Mundie, Microsoft's other co--chief technical officer: "To have scale you have to have accountability, and to have accountability you have to have numbers." Also, some Wall Street analysts noted, carving troubled businesses into separate entities could make it easier down the road to sell them or shut them down.
Financial transparency wasn't the main rationale. Ballmer wants the new structure to give the product groups a stronger voice in marketing and sales, which previously were handled centrally. To forestall the problem that had worried Gates--lack of coordination between groups whose products need to work together--Ballmer came up with a matrix-like solution called Integrated Innovation. It involves both defining who's responsible for individual technology issues and structuring the divisions to be strictly parallel. Developers in each group have exact counterparts in the other groups, with whom they are expected to stay in touch. "I'm not confused about who to go to, to get something done," says Allchin, who oversees the divisions that develop Windows Client and Windows Server products. "There's much greater clarity now."
Even Gates has gotten into the organizing act. Working with a few lieutenants, he hatched a methodology for fostering the co-development of software by multiple divisions. The result, called the Software Engineering Strategy, is a system only a geek could love. It spells out how the baton of responsibility passes from the "incubator" of an idea (perhaps a researcher in a division) to the "definer" of a product (a product-marketing guy or perhaps a division head) to the "owner" of the development team (the one who actually tells the geeks what to do). And it lays out roles for other involved parties--"participants," "reviewers," and "approver/coaches." (For example, Gates himself at various points was the incubator and definer of the Longhorn project and, once development work was underway, he shifted to mainly a reviewer or approver/coach role. Got that?) "This is the kind of process thing that Steve really gets a kick out of," jokes Gates.
Gates has always enjoyed watching Ballmer, the applied mathematician, gorge himself on corporate or administrative statistics and read the tea leaves better than anyone else. Walk down the hall at Microsoft on a given day, and you might find the CEO in a glass-walled conference room, poring over stacks of customer or employee survey results or global sales numbers. Ballmer's lust for statistics explains why he took it upon himself to revamp the actual business processes. For example, Microsoft had never defined a formal, annual process for evaluating personnel, nor was there a procedure for quarterly business review meetings for the new divisions, or for meetings considering new product or technology proposals.
Ballmer's solution was to put some of Microsoft's most popular products to work: He and his staff designed fill-in-the-blank templates for PowerPoint presentations or Excel spreadsheets for each process. "My best tool to force people to talk about something I want to hear about is to give them a PowerPoint template and say, 'Use it,' " says Ballmer. "The template lets me control the agenda." He and his aides constantly fine-tune the templates. Once he even asked the human resources department to add a field in the personnel review template for the hometown of every employee so that he could make small talk about local sports teams and the like when he knew he was going to meet an employee for the first time.
Ballmer never expected new titles and responsibilities alone to motivate Microsoft's legions; he also brought his analytical powers to bear on their pay. His abolition of stock options, which startled the high-tech world last July, cured one of the biggest headaches Microsoft faced after the dot-com crash. Though hit less hard than most tech stocks, Microsoft's shares lost half their value, putting many employees' hundreds of millions of unexercised options under water. In the future, Ballmer announced, Microsoft will reward employees solely with grants of restricted stock. And to partially recompense holders of those underwater options, he enlisted J.P. Morgan Chase in a novel deal to make a secondary market for them. It eventually paid employees an average of $1.11 per share for 345 million options. According to one of Microsoft's instant employee surveys, the move was a hit.
It's far too early to tell whether all of Ballmer's tinkering under the hood will pay off. That said, Ballmer's scheme already is making it easier for Microsoft to target competitive threats and opportunities. Allchin acted on his own to deploy a squadron of crack programmers to cook up antidotes to software viruses. (He pulled several people off Longhorn work without having to consult with Gates or Ballmer.) Group VP Raikes established a special group in his Business Services division to launch a major campaign to develop and sell complete, integrated enterprise software to small and medium-sized businesses, a market Microsoft believes is potentially far larger than the FORTUNE 500. And most of the new divisions have quickly responded to Ballmer's edict that they build automated online customer-feedback mechanisms for reporting bugs and glitches and requests for new features into all of Microsoft's software products. When a piece of software screws up, an interactive box pops on a user's screen inviting him to document the failure and notify geeks at Microsoft.
Perhaps most important, Ballmer has lifted morale in dreary Redmond. His new structure has employees like CFO Connors totally pumped: "If we get this place completely wired, man, we'll be dangerous," he says. The exodus of talent has largely stopped. Now Microsoft millionaires like Vaskevitch are sticking around because, as he puts it, "they sense a chance to do one last great thing in their working lives" by participating in Gates' grand vision of seamless computing. And as Bill himself says, "When people come to work, it's important that they be connected to a dream."
Employee fervor is crucial to fend off the threats of Linux and IBM. Integrated Innovation has to be more than a buzzword; it has to be a religion. But then Microsoft has always been something of a cult. Not so much a cult of Bill worshippers, or Ballmerites, but an army of techno-zealots who genuinely believe that they can bend and mold software to solve any problem.