Microsoft's Big Bet

Can the software giant head full steam into the Net--and still cling to Windows? It's going to try

It was 10:30 p.m. on June 21, and Microsoft Corp.'s top executives were verging on giddy as they gathered in the company's spiffy new conference center. In less than 12 hours, they would unveil .Net, the most important new strategy since the launch of Microsoft's Windows software 15 years ago. For half a decade, Microsoft had been wrestling with how to reinvent itself as an Internet titan with the same indomitable clout it held in the computer industry. Finally, the top brass had the answer: a technology that could link thousands of Web sites together in an Internet-style bucket brigade. With this new software, one mouse click would have the wallop of dozens of steps on the Web today.

Despite the high spirits, Vic Gundotra felt something was missing: drama. The 31-year-old mid-level manager had nurtured .Net since it was a gleam in the eyes of nerdy code writers. He wanted a gesture that would show everybody in the company how important this moment was. Microsoft typically hoists a new flag when a major product is launched, so Gundotra got on the phone and ordered up a vinyl .Net banner to be produced overnight. At 7:30 the next morning he rushed to the flagpoles. Without thinking, he started to lower the flag with Microsoft's Windows logo on it. Steve Cellini, another Microsoft manager, stopped him. It didn't seem right to replace the Windows flag. "We weren't that crazy," Gundotra says. Instead, he chose another flagpole and lowered the banner of Microsoft's BackOffice software.

The flag raising took just two minutes, yet it symbolizes the biggest dilemma Microsoft has faced in nearly two decades: Windows or the Internet? Windows is an amazing cash cow that has made Microsoft a $23 billion-a-year powerhouse. Yet Windows may not be crucial for businesses and consumers that use the Net to get information and conduct transactions. .Net software, on the other hand, generates zippo cash today, yet it could become a centerpiece in the new Web world. What to do? Should Microsoft stick with the business that pays its bills? Or should it put most of its brawn behind brand new technology and let Windows quietly ride out its glory days?

In classic Microsoft fashion, Chairman William H. Gates III wants to have his cake and eat it, too. He and CEO Steven A. Ballmer have decided to marry the .Net technology to Windows--a path that has sparked controversy within and without Microsoft. Even though experts say it's nearly impossible to successfully hitch a product from one technology era to a product made for the next era, Gates is steadfast that this is the right thing for Microsoft. "You have to have as much of a single strategy as possible," he says. "There are separate businesses, and there are separate competitive battles. But it's got to be within one jihad."

"DIFFERENT FIEFDOMS." Or does it? When seismic technology shifts like this occur, rarely does a company that dominates one epoch gain the upper hand in the next. Look at IBM. It clung so hard and so long to its mainframes that it never regained industry leadership after the PC revolution took hold. Microsoft could find itself similarly sidelined if it can't manage the highly delicate balance between milking the past and mining the future. "They are very, very good managers and have extraordinary resources to throw at the problem," says Harvard Business School Associate Professor Clayton M. Christensen, author of The Innovator's Dilemma, a book about just such corporate challenges. "But if the history of other industries can be used as a guide, the odds are against them."

Indeed, inside the company there are forces that are compounding the situation. On one side are managers who argue passionately that Windows should remain the top priority and that nothing should be done that would weaken its hold on the PC industry. "Microsoft is a bunch of different fiefdoms," says Brian Valentine, senior vice-president in charge of Windows. He believes that, "as .Net becomes real, the focus needs to be on keeping the PC important." While the company's Internet advocates wouldn't be caught dead dissing Windows publicly, a steady stream of executives who quit over the past two years give voice to their frustrations. They say the way to succeed on the Web is by creating a separate group dedicated to new Internet technology that need not have ties to Windows. "Can you build on Windows to get to that Internet vision? My answer is no," says a former top Microsoft Web strategist.

For their part, Gates and Ballmer reject the notion that the company stands divided. "It's not Microsoft vs. Microsoft. I don't think there are different factions. There are different points of view," says Ballmer. He says his troops are on board, though he concedes there is much to be ironed out. "The devil is in the details of how we get there," he says. Gates, too, acknowledges "there's nothing particularly easy" about the transition. He, however, argues that Microsoft has done this kind of thing before and points to the late 1980s when the company shifted from its DOS operating system to Windows. "The whole innovator's dilemma is about companies that keep on selling the same thing again and again. We don't have that," Gates says. Maybe so, but the Microsoft employees who kept working on DOS felt left behind. "It was like being assigned to Siberia," says Michael Cusumano, a professor at the Massachusetts Institute of Technology's Sloan School of Management, who believes Microsoft has the skills to move from one world to the next.

While this may not be a full-blown Hatfield and McCoy feud, the conflicting forces could be a drag on Microsoft's .Net push when it has miles to go, and backers to win. For Microsoft, following the middle path requires a massive overhaul of its software and the way it does business. The company plans on creating .Net versions of its word-processing and spreadsheet programs to run on the Web as services that people buy through subscriptions. It's crafting a new family of .Net Web services that automate the way sites link to one another. While it won't be necessary for Web-site developers or consumers to use Windows computers to take advantage of all this new technology, the company plans on making these .Net Web sites and services work best with Windows. That way, the execs hope, Microsoft's focus won't be fractured.

Microsoft will have to act fast to make all this work since its fortunes already are fading. For the past decade, it has ridden the PC explosion, ringing up an eye-popping 36% average annual growth rate. Now, even though the Windows 2000 operating system is starting to take off, PC sales growth rates are slowing. In its first quarter ended Sept. 30, Microsoft's sales rose just 7.7%. Add in the cloud of an antitrust ruling that could lead to its breakup, and Microsoft's market cap has been cut by more than half, down from a peak of $645 billion earlier this year. Worse, the way Microsoft is trying to weave Windows with .Net might land it in more trouble with the Justice Dept. down the road. Trustbusters want to snap Microsoft in two, in part to prevent the software maker from leveraging its Windows monopoly into new businesses. If Microsoft makes .Net work best with Windows, companies that build competing Web services could be aced out of the market.

Even if Microsoft sidesteps its legal problems, it still has other hurdles to clear. Several key executives who conceived the .Net plan and persuaded Gates and Ballmer to adopt it have quit in recent months. Perhaps the most troubling departure was Paul A. Maritz, a group vice-president who announced his retirement on Sept. 13--privately citing a health scare he had this year. The former Windows chief skillfully navigated the .Net plan between rival factions. In his absence, there is no obvious peacekeeper.

At the same time, the company has to woo software and Web site developers so they will adopt the .Net technology. Microsoft already has one surprising convert. Marc Andreessen, a co-founder of browser rival Netscape Communications Corp., says his new company, Loudcloud Inc., plans on using Microsoft's .Net technology for customers that hire it to manage their Web sites. "They've got the wind at their backs in terms of where the industry is going," says Andreessen. To other software developers, the notion of Microsoft playing by the Internet's open standards is unfathomable. Basic Web technologies are free to be used as anyone wishes, but Windows is proprietary technology that Microsoft guards with the ferocity of a Rottweiler. "The question that everyone asks is how open will Microsoft be?" says Keith Costello, president of e-commerce software maker Intershop Communications Inc. "There's a difference between rhetoric and reality."

POWER CLICKS. If Microsoft pulls this off, the software maker could utterly transform the Internet. The company aims to provide technology that will weave the Web into a more powerful fabric, making it much easier for people to get things done. At the core of the .Net plan is a powerful idea: letting unrelated Web sites talk to one another and to programs that run on every device--from PCs to TV set-top boxes. When that happens, one click of the mouse could set off a cascade of actions without the computer user having to visit additional Web sites or open new programs. For instance, if a business's purchasing agent wants to get the best price for a ream of copy paper, she'll be able to comparison-shop from multiple suppliers at one Web site, instead of having to click from one site to another. Because .Net promises to let different computer systems talk to one another, once the purchasing agent clicks her mouse to buy the paper, it's automatically logged into her company's accounting and inventory systems. Today, that's generally handled manually.

If Microsoft's .Net strategy fails, it will be stuck in the past--another example of the difficulty of corporate reinvention. To understand just how difficult this transformation will be, consider the tortuous path that Gates, Ballmer, Gundotra, Maritz, and dozens of other Microsoft execs have taken to come this far. This tale was pieced together from more than 75 interviews with present and former Microsoft managers, partners, and competitors, some of whom would speak only anonymously.

NO TURNING BACK. The .Net strategy emerged from the ashes of a supersecret Internet initiative that went down in flames last year. Called Project 42 and headed by J. Allard, one of the famed architects of Microsoft's Web browser strategy, the group's ambitious goal was to create tools that corporations and consumer Web operators could use to construct high-volume e-businesses. Project 42 got its name from the book The Hitchhiker's Guide to the Galaxy, the cult classic in which the earth is really just a lab experiment run by mice to find the answer to the question: What's the meaning of the universe? The answer they find: 42--that, and nothing more. Project 42's purpose turned out to be just as feckless. After a soulsearching exercise in May, 1999, Microsoft's chiefs disbanded the organization, which had bloated to an unwieldy 1,500 people. "It was incredibly ambitious and naive," says Operating Systems boss James E. Allchin. "As soon as you say you're going to solve every problem, it's a plan for failure."

Yet there was no turning back on the Web. It was clear to a handful of Microsoft's bigs that the company had to have a bigger Internet presence. The more that software programs run on the Web, the less important Windows becomes. The impetus to get with the Web came from an unlikely source--Maritz. He had helped quell an insurgent group in 1997 that urged the company to build a Web software platform that could have competed with Windows. Ironically, Maritz' ties to the Windows world ultimately made him the best candidate to lead the Internet charge. The 45-year-old Zimbabwean could persuade others to bet on what ultimately became known as .Net precisely because he so clearly wanted Windows to succeed as well. Maritz, a 14-year veteran of Microsoft's top ranks who had never gotten much credit, had an opportunity to put his mark on one of the company's boldest strategies yet. He asked to be put in charge of the company's software developer tools division--what looked like a demotion. While his job seemed to shrink, he was thinking big.

To get his team going, Maritz pulled a page out of the Microsoft management playbook. He scheduled a daylong event to give developers a road map of Microsoft's latest Web strategy. That meant he and his crew had a deadline for coming up with a plan and selling it to the higher-ups. "Nothing focuses the mind like the sight of the gallows," Maritz would say to Tod Nielsen, a former Microsoft vice-president and Maritz' second-in-command at the time. It was late spring, 1999, and Maritz' band of about a dozen rebels had until Sept. 13 to get its act together.

While they raced against the clock, conflicts emerged between the old Microsoft and the new. The company was only a few months from shipping a test version of SQL Server 2000, its database software. A small group of developers approached Paul Flessner, the senior vice-president who runs the business, and asked him to delay the product release a few months so they could deeply weave in technology, known as XML, a programming language that creates tighter links between products, prices, and other information on the Web.

Since then, XML has become a crucial ingredient of .Net. But back then, Flessner knew the delay to SQL--a $1 billion-a-year product--would be hugely expensive. He stewed. A few days later, when he visited customers in the Midwest, they were asking about XML. He gathered up his courage, and on May 1, he met with Gates and other top execs to ask for a delay. "There was a lot of teeth-gnashing," Flessner recalls. They worried about lost revenue. Flessner persuaded the bosses that SQL would have to wait. It was shipped this past September--nine months late, but with XML.

Time and again, Microsoft kept bumping into itself. It turns out that one of the people who needed the most convincing was Ballmer. He had agreed to give the keynote speech for the September Web strategy event. But the night before, when he arrived at the ballroom of the Sheraton Palace Hotel in San Francisco to rehearse his speech, he was, well, underwhelmed. Maritz' group had put together a presentation for him that explained how developers could use Microsoft software to build Web sites that could share information with one another. "The speech didn't seem to be bigger than the sum of its parts," Ballmer says.

He grilled them and they rewrote the speech over and over. Ballmer got them to add a slide about Passport, a feature of MSN that would appear a month later. It has since emerged as the prototypical .Net service. Passport lets Netizens store their shipping and credit-card information on the Passport site. When Netizens shop at e-tailers who use Passport technology, the billing and shipping information automatically pops up. That way, people don't have to type in the info anew on each different site. Still, Ballmer wasn't satisfied. In the wee hours, he took aside Nielsen, the highest-ranking executive in the developer group who was there. Nielsen is a cheerful man who refers to himself sometimes as "the Todster," but there was no joy in Todsterville that night. "I just remember him saying: `Tod, you let me down,"' Nielsen says. He was devastated. Minutes later, Ballmer got in the elevator and went to his room. He was too pumped up to sleep, though, so he turned on ESPN, watched SportsCenter--and mulled over this new Web software.

By morning, Ballmer saw the bigger picture. "It's fair to categorize it as an epiphany," he says. "It was really that night that I got my opportunity to figure out how the pieces fit together." When he delivered his speech, he felt comfortable with the idea that services would replace applications as the way that people use software. And he could see how all the services on Microsoft's online service MSN could benefit from the new programming tools Maritz' crew was creating.

By the time Ballmer addressed Microsoft's annual employee meeting three days later, he had become a raving Net evangelist. Arriving to the theme music from Disney's Tarzan, Ballmer stormed the stage, lip-syncing the song. He bounced into the audience, high-fiving dozens of the 14,000 people gathered at Seattle's Kingdome. By the time he started to speak, he was dripping with sweat. His message: Microsoft had a new "it." The company would throw its weight behind the move away from traditional software applications and toward software delivered as services over the Web. "This is going to be a way of life," Ballmer told the crowd. But Ballmer made one point very clear. Microsoft wasn't giving up on the PC and Windows. "The old `it' doesn't go away," he said. "We're going to be a two-`it' company for a while." The army had its marching orders. Build new technology for the Internet. But build it on the back of Windows.

With the big ideas sketched out, Martiz' crew had to fill in the details. Just how would Web sites tap Windows? A turning point came when Charles Fitzgerald, head of business development for Maritz' group, came upon a clever Web service called Ltd. It helps online shoppers find the best deal on the Web. What's novel about RUSure is the tiny bit of code that shoppers download from the Web to their personal computers. The code works hand-in-hand with the Web site, so that when RUSure users shop for the latest John Grisham novel at Inc., for example, they are automatically notified of better deals for the book from other e-tailers. It was eye-opening for Fitzgerald. Here was a nifty little Web service that took advantage of the power of a Windows PC and the ability of the Web to update itself constantly. "Microsoft wins when we change the rules of the game," Fitzgerald says.

The pieces were falling into place. It was time to show Gates. In October, Maritz' group paraded into Microsoft's boardroom just outside Gates's office to reveal their work to the company's Technology Leadership Team. The group included Gates, Ballmer, Maritz, then-Office chief Robert L. Muglia, Allchin, and newly hired Internet czar Richard E. Belluzzo. It seemed like a simple demonstration. The group had created a small e-mail program that ran on the Web but worked better when it interacted with software running on a PC. It was the first step on the long road to .Net.

But Gates quickly lost his patience. He thought the plan required information to be shot off to the Web with each keystroke, slowing things to a crawl. Consumers would never stand for it. He pounded the table. Maritz calmed him down: "Bill, you're really going to have to stop and listen." The group showed how the Web program was even snappier when it interacted with Windows because it could tap into the PC to store information--making it readily accessible. Now Gates was sold. "It was the first time I was willing to say, `Okay, let's center the strategy around this,"' Gates says.

SUBSCRIPTIONS, ANYONE? Gates now thought he had the answer to a problem that had vexed the company for years: how to make Windows crucial in the Internet Age. To Gates, the only business that could work for Microsoft was one that represented what he calls "a singularity." In science, a singularity is something that is distinct. A black hole, for example, is a singularity. Gates wants products that--like a black hole--have such an intense gravitational field that they draw others to use them. The more people used it, the more powerful it became. .Net "isn't just an addition to the platform. It is the platform," Gates says, because Windows and .Net technologies are intertwined.

But how would Microsoft make money from .Net? Microsoft would have to .Net-ize its existing products, weaving the new technology into its operating systems and applications so they're ready to work in the new world. But it wouldn't get paid extra for that. Gates wanted to know where the money would come from. "We're not a zero-revenue company," he is said to have thundered to Maritz. At first, some of the execs thought the answer was advertising. The company would offer online versions of its Office applications, for instance, and slap advertising on the Web site to generate revenue. Gates was dubious. And, once Ballmer did some research, it was clear that advertising wouldn't pay the bills.

Instead, Microsoft hopes consumers will subscribe to the online versions of its products, such as Office.Net. Gates envisions four flavors: a free bare-bones version, and versions with increasing functionality that could cost from $3 to $19 a month. At the low end of Office.Net, for example, you could get basic word processing and a minimum amount of online storage. Pay top dollar, and get the full word-processing and spreadsheet package, gobs of storage space, and online support to boot. Tie that in with other .Net services, such as scheduling, and users could buy a plane ticket online and have the flight schedule automatically show up in their calendar in Office. Microsoft plans to charge for other services, too. Passport is free to consumers, but Microsoft plans to make it one of a group of services that it hopes to collect fees for.

For the next nine months, the struggle would be over how exactly the new "it" would blend with the old. "We've had more top management meetings--the top five or six people sitting in a room for three hours straight--since the beginning of the year than ever in the history of the company," Gates says. He gave the Net strategy a huge boost in January, when he handed over CEO duties to Ballmer and took the title of chief software architect. The chairman wanted to spend more time developing the new strategy. And he figured things wouldn't move quickly enough unless he took up the campaign himself.

Picking a name, even a code name, for the technology turned out to be tricky. In fact, the struggle over selecting a name turned out to be instrumental in coming to terms with the strategy itself. The name had to say "new." Initially Microsoft couldn't back away from Windows. Ballmer suggested Next Generation Services, but Gates quickly shot that down. "I said, `No, no, no. You've got to put Windows in it,"' Gates says. So the name became Next Generation Windows Services. The battle wasn't over, though. Maritz was one who wanted to drop Windows from the name. Using Windows would have held the new strategy forever captive to the Microsoft of the past. It represented old thinking.

Ultimately, the entire Microsoft executive brain trust got dragged in. Before they were done with the months-long process they considered more than 75 names, including ICE, for Integrated Computing Experience. Gates's favorite was AAAware-as in "anyplace, anytime, anywhere" software. But he was overruled by arguments that AAAware Office or AAAware Windows didn't exactly roll off the tongue. Just a week before launching the strategy, the group chose .Net. The tag could easily be added to Microsoft's existing product line. "There was a way to convey the you-can-have-your-cake-and-eat-it-too strategy," says Craig Mundie, a senior vice-president for consumer strategy.

Now, the challenge for Microsoft is trying to do what's right for Windows and Office without hampering its Web projects--and vice versa. So far, about 3,700 of Microsoft's 40,000 employees are working exclusively on .Net. By the end of next year, though, Gates expects virtually all of the company to be focused on creating new .Net products or updating old products with the new technology. He has already committed to spending half of the company's $4 billion annual research-and-development budget on .Net breakthroughs.

It's unlikely that the internal transition will go smoothly. New skirmishes already are breaking out between .Net apostles and those who run Microsoft's traditional businesses. One example: NetDocs vs. Office. Microsoft manager Brian MacDonald leads a group that's creating a new Web word-processing and calendaring system called NetDocs. "Brian would like to roll out NetDocs aggressively," says Eric Rudder, a Microsoft vice-president. But MacDonald is meeting resistance from Steven Sinofsky, who runs the nearly $10 billion-a-year Office business. It's unclear when NetDocs will see the light of day. Sinofsky and MacDonald decline to comment, and Muglia, who oversees Microsoft's .Net services, says, "How it emerges as a product remains to be seen."

DEFYING HISTORY. Ballmer says such squabbles are healthy. The company needs workers who "get passionate about what they're doing." If they butt heads with colleagues, all the better, since it will ultimately produce better products. NetDocs and Office will march down separate paths--NetDocs for consumers and Office for corporate customers.

Even while those turf battles play out inside Microsoft, some of the company's top Internet thinkers have quit rather than try to meld their new ideas into its A-to-Z strategy. Prime example: the folks who started Crossgain Corp., a software company that's helping programmers to interlink Web sites. It's like a piece of .Net--only without the tie-ins to Windows. Worse yet for Microsoft, Crossgain's leaders came from the software giant's Internet A-Team. Chairman Adam Bosworth led Microsoft's first XML efforts. The CEO is Tod Nielsen. Another former Microsoftee, Thomas Reardon, who led its first Web browser team, quit to be CEO of Avogadro Inc., a wireless Web company. "The grand unification attempts at Microsoft became exhausting," says Reardon. "They do big really well, but it's very difficult to exploit niches."

Microsoft thinks it can defy history and serve two masters. But with IBM's struggles to remain an industry leader serving as a warning, Microsoft's leaders are anything but overconfident. Even the chief architect of .Net recognizes the dilemma. "There are threats to existing businesses, but there are also new opportunities," Maritz says. "How those are all going to balance out, we're still not sure." As Microsoft encourages developers to create .Net applications for the Web, Windows could be left behind. For now, though, only one of Microsoft's two "its" pays the bills. And there are no plans to lower the Windows flag anytime soon.

Before it's here, it's on the Bloomberg Terminal.