The Four Hundred

Microsoft held its annual financial analysts conference in Redmond, Washington, last week, and the top brass of the company spent 10 solid hours with Wall Street and IT analysts trying to explain what .NET is, why it matters, and how Microsoft is going to make money. The company also identified its three biggest obstacles: IBM, Linux, and Java; it even mentioned the AS/400 twice in discussing its marketing plans.

In his opening talk at the conference, Microsoft's chairman and chief software architect, Bill Gates, said that while the software industry was the victim of a big crunch on spending, Microsoft was going to increase its research and development spending, which mostly goes to salaries, by more than 20 percent in fiscal 2003 (which ends next June), up to $5.2 billion, from $4.2 billion in fiscal 2002. Part of that money, we later learned, is going for pay raises, and the remainder is going for new hires. Microsoft CEO Steve Ballmer said the company added close to 5,900 employees in fiscal 2002 and that it would add 7,500 employees in fiscal 2003. Last year, after attrition, Microsoft had a net gain of about 3,000 employees, for a total of 50,000 worldwide, and given current attrition rates, it will grow by about 5,000 employees in 2003. "That's an over 20 percent increase that, at least if you're looking at this in the short term, is very counter-intuitive," explained Gates. "You might ask 'Why don't you cut it 20 percent?' Well, we have a lot of belief in the things we're doing, in terms of growing the markets we're in, gaining share in the markets we're in, and doing a lot of things that will help this industry be very strong in the years ahead. So it's an aggressive approach, but one that I certainly believe in."

While no one would be stupid enough to argue that Microsoft should not aggressively recruit college students and other hot IT talent to bolster its position--something the company can do, and has done, as there has been a flight to solid companies since the dot-com collapse--very few companies can afford to be so generous or aggressive in a down economy--or, indeed, in any economy. Every time a Microsoft executive talks about spending a billion dollars here or a billion dollars there, you have to remember that Microsoft had $38.6 billion in cash and short-term investments on hand as of June 30, an increase of over $7 billion in a year when the company only grew revenues a little more than $3 billion, to $28.4 billion. Those numbers are not typos. Microsoft has cash that is equivalent to more than 16 months of revenue in the bank, not counting accounts receivable. It's nice to have a monopoly, if you can get one.

Microsoft's spending on research and development, and on marketing or any other expense, is not limited in the way that other companies' spending is. Look at IBM, for example. Big Blue had $3.6 billion in cash and marketable securities on hand on June 30, and it has just spent $2.7 billion of that on acquiring the consulting business from its auditor, PricewaterhouseCoopers. IBM, in sharp contrast to Microsoft, had cash on hand that was equivalent to 15 days of revenue, based on its sales rate of nearly $86 billion in 2001. (The PwC business IBM is buying should fill in the revenue gap from the disk business that IBM is exiting, and the cash generated from the disk sale should refill the coffers after the PwC acquisition, too.)

Microsoft's pile of cash is one reason why the company, as Paul Flessner, senior vice president of .NET Enterprise Servers, explained, can increase its server sales people by 22 percent, its account managers by 15 percent, and its sales people dedicated to developer tools by 47 percent in fiscal 2003 without worrying about its ability to pay for such expansion. Microsoft also plans to spend $200 million on an advertising campaign focused on promoting the lower total-cost-of-ownership of Microsoft technologies, compared with alternatives like Unix servers, and another $500 million in new channel investments. Flessner said that in the server market alone, Microsoft would dedicate more than $2 billion of its 2003 budget to beefing up its server programs, which is necessary so they can take on Unix and other proprietary alternatives like IBM mainframe and AS/400 servers. The TCO talk and integrated platform pitch is right out of the AS/400 playbook, which IBM has been using for 14 years.

In any event, a cynic would say that Microsoft is spending more money in fiscal 2003 on R&D because the next two generations of Windows--the "Longhorn" and "Blackcomb" releases of its operating systems--are going to be years later than expected, and it needs to throw bodies at the problem. But this is not cynical enough, actually. These code names are nearly meaningless, since Microsoft keeps changing what it plans on doing as competitive conditions and innovations occur in the IT market. In talking about the future .NET products, Gates referred to three waves of innovation. The current wave of products is most accurately defined as XML-enabled and SOAP-enabled Visual Studio.NET, announced in February, with its Common Language Runtime environment and C# Java-like programming language, and the forthcoming Windows .NET Server operating system, which as we report elsewhere in this issue, has just come out in its first release candidate form. Further down the road, in what Gates called the second wave, the "Yukon" release of its SQL Server database, which has an XML-friendly file system that will eventually show up in wave three, the Longhorn generation, which will also see a big XML-based release of Office.

"This is a very big deal," explained Gates. "We don't have some exact schedule for this. In fact, because it's very technology driven over the next several years, we're going to be working hard on the definitions of this and know exactly when it gets pulled together." That, in plain English, means Microsoft doesn't know when any of this stuff will be available because what it is trying to accomplish is difficult and because it is in the middle of radically revamping the process that its thousands of software engineers use to create programs, and there is no way to quantify rollout schedules until Microsoft gets a handle on all of this. If Microsoft spent over $100 million in two months looking for serious bugs in Windows as part of its Trustworthy Computing initiative, I have no problem believing that Microsoft will have to spend billions of dollars and several years cleaning up what it has done so far with .NET, as it rolls new features into .NET. Microsoft will never offer this as an explanation for its actions, but it just makes good sense.

As a wise man--Frank Soltis, the chief architect of the OS/400 architecture over at IBM--once said (and I am paraphrasing here), it takes five years to release a truly new operating system, and there is no amount of money or bodies that you can throw at the problem to change this. You can do one-fifth of the new features every year and call them operating system releases or operating system versions or whatever, but it takes five years. Windows NT was started in 1991, announced, some would say prematurely, in 1993, and wasn't really useable until 1995, with Windows NT 4.0. Windows 2000 took another five years to get to market and to be useable. And Longhorn, by which I mean a truly different Windows operating system, not this elusive code-name that may or may not mean something, is not going to get here until much before 2005, if Soltis and his interpretation of history is any guide.

Gates, Ballmer, and Flessner were pressed, as always, to identify their biggest competitive issues in the coming years, and while the executives waffled a bit, they all identified IBM and Linux as the two things they were gunning for. Ballmer threw in Java as a third target.

"If we look at the last year, we'd say that, competitively, we are stronger versus all of our competitors, with two exceptions," said Gates. "I would say that we are either equal to or have things to focus on, in terms of both IBM and the Linux-type software community." Gates said that Microsoft was much stronger than IBM in messaging, with Exchange outselling Domino/Notes last year, and cast an envious eye on IBM's Global Services unit and conceded that WebSphere was real. "IBM Global Services is a strong asset there, and they've managed to get WebSphere out. And they're a long-term competitor. We've enjoyed competing with them. But I'd say they've also done a lot of good things this year."

Flessner echoed this opinion later in the day, when lining up Microsoft's software stack against an IBM WebSphere-Linux stack, which Microsoft would like to characterize as the stack that IBM is peddling (it's more complicated than this, as we all know). "I heard it said best the other day," Flessner said. "Linux is free like a puppy. You know, you look at it at first, it looks pretty free, but by the time you get all the pieces on it to make it work, it's not so free. Linux is just that little piece down below, in terms of the operating system kernel. Sure, it does task management and flopping and I/O and memory. But there's a heck of a lot more to an operating system, or at least what a customer thinks of as an operating system, than just Linux. There's a lot of pieces you've got to add. Certainly, WebSphere can round those out, but WebSphere is not free." Later on in his talk, after reviewing Microsoft's plans for fiscal 2003, Flessner returned again to IBM and Linux. "IBM and Linux together are a formidable competitor to Microsoft; it's not just Linux alone, and it's not just IBM alone. I think we know how to compete with IBM. We build software. I think we build software better than IBM, and we'll do a good job over time integrating it and making it more tightly integrated. That really is the value of Bill's position, and the senior technical leadership team, all the things that we do to really tightly integrate around a single message. We're not confused about that, and we know that's an important part of what we have to do to best these two competitors."

Ballmer touched on the same issues when he gave his pep talk about how Microsoft was changing its culture, and was asked in Q&A session to identify the competitors it would kill if it had a silver bullet. "In servers and tools, I would say certainly I see a lot of opportunities. The fact that the market is so focused in the bad economy on cost plays exactly to an advantage that we have versus the incumbent providers of server software, like IBM and Sun and others. Not just the software, it's the hardware. The competition here is intense; the competition from Linux, the competition from IBM, and those really are separate. IBM may use Linux, but it's also an independent competitive force with WebSphere. J2EE, market inertia, sometimes the easiest thing to do is to just not change an application. In academic circles right now, Java, Linux, there's a lot of momentum behind those things; and certainly for all of you who represent Wall Street firms, one of the issues I know we face as we try to bring our products to market is a lot of kids coming out of school having more Java expertise than they would expertise programming on our toolset, and we need to get after that in the academic arena. The most important competitors, I guess, are the ones that are at the core. You can go outside of the wheel of businesses, but it's the stuff that we compete with on the platform sense that is probably the most significant, and that's IBM, that's Linux, that's J2EE Java. It's those phenomena. I don't know if I'd pick any one, but at the end of the day all Microsoft people are somehow competing at the platform level, no matter what business you're in. And then each of the businesses may have its own independent competitors. But we all sort of drive, or fail to drive, based on our platform competition. So that's where I guess I'd mark the silver bullet."