The business of Microsoft

The business of Microsoft
Alex Lash

Staff Writer, CNET NEWS.COM

If Bill Gates is the brains of Microsoft, then Steve Ballmer must be the brawn. Whenever the company needs a fire-breathing, podium-pounding alter ego to Gates's detached corporate coolness, the No. 2 man at Microsoft gets the call.

It is a role that has grown out of a

Sign: Aries

Originally from: Detroit (Dad works for Ford)

Hobbies: Basketball, running at 5 a.m.

Claim to fame: First business manager hired by Microsoft

Key connection: Had Bill Gates as housemate at Harvard

Janet Reno, IE bugs, and Apple

If you say who are our friends, that might be another way to get a cut through it. There are our friends in the hardware community. We have Intel, the OEMs, the IHVs [independent hardware vendors]; we have friends who help serve large enterprises either with applications, Internet access and services, or systems integration services. And that's where you get the people like Digital, British Telecom, Unisys, and HP. It turns out as we've really gone after business ISVs, there's a lot more ISVs out there than I guess we even were dreaming about. In small business again, application developers, small ISPs. Everybody knows their favorite small ISPs. Then there are a bunch of guys we call VAPs [value-added providers]. There's about 175,000 small companies that provide computer services to small business. We're actively courting them.

For the consumer market, there are retailers--people like AOL. AOL looks like a partner to us in terms of at least the browser part of our business going after the consumer space.

So friends, competitors, technology areas, products, customers, visions...I think it paints a pretty broad picture, which is simply to state the affairs of Microsoft today.

One competitor that you didn't mention was the DOJ. Let's get that out of the way. You characterized Janet Reno as a joke?
Never. I, to my chagrin, said "To heck with Janet Reno" when asked whether she would try to stop us from shipping Small Business Server because it's such a wonderfully integrated product. I said, "If she did, then the heck with her because this is the American way." No, I don't consider the Justice Department a competitor. I think what we're doing is right and lawful and moral and proper and competitive. I might even say it's the American way. We're innovating, we're adding value, we're driving down prices, we're competing, we're serving our customers, and we're doing it well. A lot of other companies in the United States are benefiting because they're building on top of our platform and thriving! So I might start playing "The Star-Spangled Banner" if I went too long!

So I believe--and I believe in the government and I think it has got to do its job--that this is a misguided investigation because what we've done is right, but nonetheless, I would never call the Department of Justice a competitor. They're trying to do their jobs even if I don't think this is particularly an area in which they have a case.

Is there a backup plan in case the Justice Department prevails?
No, no backup plan.

Why not? You contribute $2.6 billion in R&D, so it's got to go toward...
Who knows what will happen? Who knows how long it will take to happen? Who knows, who knows, who knows? Our job has got to be to keep focusing on the customer and serving the customer and not sort of just wasting people's time and energy trying to anticipate what a judge might decide someday. We'll cross that bridge if we come to it.

The follow-up question is, is this your hope or intent that [the DOJ investigation] is delayed at least until Windows 98 comes out and then the issue becomes a moot point?
No, I can't say it's my hope or intent. I'm not involved day-to-day in the legal strategy, but it is our hope and expectation that the judge will concur with us, that what we've done is right and proper and innovative and procompetitive because it serves the consumer very well. We hope the judge will agree with us. We expect the judge to agree with us that we are in charge of trying to serve our customers and deciding what goes into our products. Our customers can decide whether to buy them or not buy them as they see fit. We expect the judge will agree with us that we have a right to sign contracts with our partners that say they cannot convolute our product in the distribution process that they can't take pieces of our product out.

When is Windows 98 coming out?
I think we're still saying second quarter of next year.

Is all this bad publicity making customers think of options other than Microsoft?
I think at the end of the day customers say, "What product helps me get my job done? What product helps me help my company get their job done? What's going to help me at home?" And they buy what's going to take care of their need. They read with interest about the DOJ suit, but then they focus on what serves them best. Our products are still doing a pretty good job of that.

Why did you ship IE 4 with its level of bugs and performance problems?
[Laughing] "No, I stopped beating my wife!!" We shipped IE when the IE 4 team--using its best, conscientious, professional judgement--thought it was ready. Does it have some things that could stand improvement? Yeah, it has some things that can and are being improved. There's no doubt about that.

I don't think the team believes now nor believed then that they shipped prematurely, particularly relative to how they perceived the expectation set among the leading-edge customers who use their product. In some senses, you do have different expectation sets for different customer types at different times.

I'm not trying to say there's some excuse that we should ship bugs in anything ever at any time, but the reality is you're always making a set of trade-offs about probability of problem, unknown problem, vs. when you ship. So you're always going to get some problem that occurs after you ship or some number of problems, and you make your best trade-off. That trade-off isn't made under some absolute rule set. What is right for the Age of Empires [a Microsoft game] is not the same as what's right for IE; it's not the same as what's right for NT Server. I think the guys think they made a right set of trade-offs and I think they know they're doing some more work to continue to improve that product. I know they would argue to this day that they certainly had a least the quality level of their competition at the time their competition shipped. You might give me feedback to the contrary...

We'd totally disagree.
You totally disagree?

Yes. Both products have bugs, maybe the same number of bugs, but the bugs in IE are system-level; bugs in Netscape Navigator are application-level. So if Netscape crashes, it crashes the app; if IE crashes, it screws up your install, your Windows install. Big difference.
I will certainly share that feedback and they can think about it and they may try to rebut.

When you're talking about friends and competitors, one company that didn't show up on either side was Apple. Where do they fit in?
They're friendly. I wouldn't list them quite at the same level of some of the guys that I chose to list in the sense of "do they help our business and do we help their business in the same way we and Digital help each other in enterprise services." It's not that critical of a relationship either to Apple or to Microsoft. I'm not saying it's a small thing, but it's not a day-to-day active thing in quite the way the Digital relationship is.

So you think that what, the quarter-billion dollars or whatever you make off Apple sales will continue or is it going to be a smaller business?
It is a business that has been shrinking in size over the last several years. It has not been a growth business for us.

I also didn't hear you mention any friends or competitors, except perhaps AOL, in the media space.
Banks are friends. We have a joint venture with First Data Corporation to provide technology that banks can use for bill payment and bill presentment. I don't know exactly who we did announce with, but we announced a couple of banks that will join with us. I think you'll see more and more banks.

American Express is a great partner for us in the travel business. Reynolds & Reynolds is a great partner for us in the car business. NBC is obviously a unique and good partner for us in the news business.

But those businesses are so big. There's so much happening in media and content and services that I think there's plenty of room for lots of different players. I don't think anybody is ever going to get 70 percent of the ad budget of U.S. advertisers. So in some sense, everybody is a competitor and nobody is a competitor. It's not like operating systems or mail systems, where I think we can all fairly agree it would be unusual to see 30, all with great popularity. Those tend to neck down to a couple or three players much more quickly than some of these other areas.

NEXT: On Java and NCs

On Java and NCs

There are a couple of server-side questions. Some of the companies that you've mentioned as competitors are getting together and backing the enterprise JavaBean specification. What is your view on the server side of Java application development in a central IT environment...How much of a threat do you see that being to NT's growth, etc.?

Without meaning to sound disingenuous, what do you mean by "Java development?" Do you mean using the Java language to write server-side code?

Yes.
Really, yes?

Yes.
Well, the language needs to be richer; otherwise, you've always got to remember to throw in sort of an extra word. At the server level, what's the goal? Can people write server-side programs in Java? Yeah, we believe in that. If the [issue] is should people only use Java run-time services to write server applications, I think that's pretty naive. Are you really not going to take advantage of Unix security when you're writing a Java security application? Are you only going to use Java security? Are you only going to use the Java threading model, which is lousy today, for server-side applications? You're not going to use anything in Unix? (Forget NT here for a second. All the same arguments apply.) Are you not going to use the admin model? Are you not going to use the directory model? If you're not going to use any of that stuff, why even bother hosting on Unix? Why not just throw in a couple of device drivers and run naked to the hardware?

The fact of the matter is the only sensible way to write a server-side program in Java is to take advantage of Unix system services. You might use some of the Java classes or not; that's up to you. But darn, you cannot do something that's good and competitive without using some of the native services of the platform. That's true whether it's Unix; that's true whether it's NT. So do we actually think people shouldn't write server-side code in Java? Well, people can write server-side code in Java, but I think people who want to do good work will take advantage of Windows NT and Unix--just to take two examples, Windows NT and Unix server-side services.

So I don't think this is about "write once, run anywhere," and I don't think it's about the popularity of the Java classes as a platform. That may be the intent of the authors, but that's not going to happen.

You said that one of Microsoft's ten goals was to increase simplicity and manageability. This scenario that you're talking about doesn't really seem to make things simpler for developers. What would make things simpler for developers would be a "write once, run anywhere" situation...
Java is hopeless for developers today, hopeless! It doesn't have enough services to be interesting really for developers today. You start with something that doesn't do much called the Java classes. They just don't do much! One way to make things better for developers would be to write a whole new operating system that ran on all other operating systems so eventually those other OSes would be obsolete. That is one way to make life simpler for developers.

Another--take it from the mouths of the guys who are in charge of Windows--would simply be to make Windows better, so they need to put Windows everywhere so that Windows became the place you wanted to do all those things! The honest truth of the matter is it's a long process, no matter which approach you take. The whole notion of "write once, run anywhere," except for maybe some small applications--small, small, small, no significant applications--it's not likely be even something you can dream about sensibly for a number of years. For small applications, some maybe, maybe, maybe...All I'll concede is maybe.

On the other hand, are there nice things in the Java programming languages? Are there nice things about its object model, its typing model? Yeah, there are some nice things about that, and we're trying to make sure that we put those in other languages and we bring those generally to the Windows platform. But the notion that says the only way to do printing is "the Java way..." Oh, sorry! I forgot--Java doesn't do printing. Sorry, cheap shot. It's a cheap shot, but it is part of the core point here!

What are your projections for the Hydra market versus the NC market?
I actually think a lot of what Hydra is going to do is help down-level PCs access up-level applications. From today's model, I'd say from a Win3.1 system you'll get a terminal emulation session to a modern application, a Win95 app.

But if you ask me about sort of Windows terminals...I don't mean to sound too definitional, but define an NC for me.

A Java-based platform.
That doesn't run Windows?

Yes.
OK, good. Then we have a common definition. I was with some analysts yesterday who would not necessarily have agreed with that definition.

I think Windows terminals, let's say four years from now...might make up five to ten percent of, let's say, all new terminal-style devices that go in. I think NCs might make up zero percent, one percent. One percent is like zero because it will mean there's no critical mass.

The key question you have to ask is why would somebody want those devices? They're incompatible, they're not small, they're not cheap. If you really just want a browser-based device, by that time I'd tell you you'd just be running a locked-down Windows station that only runs a browser, including Java VM or whatever else you want to put on it. But why? That will be cheaper because of the economics of the PC industry. That's going to be cheaper than this new, incompatible device. So I think the NC, as you defined it and as I would agree with the definition, will have a very hard time hitting critical mass. Some may sell, so when I say zero percent I'm only trying to imply it will be a rounding error.

I'm less worried about NCs as a basis for Windows competition; I'm more worried about Java as a basis for Windows competition than I was a year ago. That is this Java operating system, as much as I jabbed at it, we have to beat it. If we don't do certain things, it can overtake Windows. But I actually think it has more of a chance to become a serious competitive threat for us than the NC. The NC, I think, has lost a lot of momentum in the last year, partly because there's now Windows terminals that people can buy.

Over in the eminent horizon, we're that much closer to NT 5. And PC prices have come down. The fact that PC prices have come down is a constant source of pressure on an NC-like concept.

So I'm not saying I'm dismissing it or that I'm not going to pay any attention to it. We have some guys that just watch, looking for NC design wins and try to fight them in any way we can, calling low, calling high, calling middle, calling, calling...But today I think the Java thing is actually a bigger threat than the NC thing.

NEXT: Where does that R&D money go?

Where does that R&D money go?

Given the NT's growth curve, I'm curious if you have any worries going forward about the Justice Department looking at your server-side software operations, given the sort of all-encompassing nature of BackOffice.
I think what we are doing delivers incredible value to customers--and that's what we're supposed to do! This is America, for God's sake! We're supposed to work hard, build software, and try to offer it at a good price. So I trust and I have faith that this beloved practice of trying to serve our customers and help create a platform that helps other companies success--most of them American, I might point out--continues to succeed in the software business. What am I supposed to believe? My unfortunate quote about Small Business Server integration and the Justice Department's view of that aside, it's just a good thing! So I'm not going to go down the hypothetical path.

We are going to do a lot of R&D. The way I explain this to customers...You know I said we're in four businesses--let's just talk about the first three. I tell them there's $2 billion in R&D at Microsoft. And that $2 billion, do you know where it's going? It's going in three boxes: the Windows box, the Office box, and the BackOffice box. So unless we're incompetent every year--not physically, but sort of bitwise--every year there's more bits! More bits and more bits going into three boxes! We're just pushing in bits worth of innovation and we're going to keep them the same, more or less--everyday low prices. Some will go up, some will go down. But that's our job. Our job is to put $2 billion more of intellectual capital into three boxes every year. That's good for the customer.

So you ask me, are we going to keep doing more integration, more value-add, more R&D, extending the boundaries of what it means to be Windows or Office or BackOffice? Yes, I'm saying that's our strategy. It's good for the customer and it's good for software companies, most of them. There are some it's not good for.

Believe me, it's not good for some guys who used to build TCP/IP stacks that we put one into Windows. But is anyone really going to argue that it was bad for the advance of the software industry to have a TCP/IP stack built into Windows? That's the kind of discussion this is.

When does NT 5.0 come out?
Now you're unfair. The other questions were all super fair! I'd say roughly a year from now, roughly. Who here has looked at the beta?

We haven't.
There you go. That will tell you why the date is pushed back. It's not code-complete and it's buggy, and we've got a lot of work to do!

NT 5, I think, will be at least for medium and large companies a bigger milestone than Win95 was. There isn't going to be any Rolling Stones' "Start Me Up," blah, blah, blah. But NT 5 itself, Exchange on NT 5, SQL Server on NT 5, our new version of Office--I think there will be a major corporate migration designed to improve digital nervous systems and reduce costs. So it's absolutely essential we get that product right. And if that is a few months later...I hate it, I'd love to have it now, but we're better off getting it right.

So it seems like one less reason for Windows 98 to go into corporations.
The focus for business is NT Workstation. There's no confusion.

NTW 5?
NTW 4, I would say today. I recommend to customers on new computers today that they should take NT 4. I'm not really trying to tell them that they should go throw out a bunch of machines or upgrade a bunch of existing machines from whatever it is to NT 4. But today, I would say it is a smart thing to bring in NT 4 on new machines in business. At the time of NT 5, I think actually people would be smart to at least think about what their overall installed base migration plan could look like to NT 5. I'd say it that way.

The Java dispute with Sun and Microsoft, is that going to be resolved out of court, in court?
Based on my last brief, I'd be surprised if it was resolved out of court.

What is it that you fear from Netscape?
A way to compete with Windows, let's call it the Java way and the Netscape way, which are sort of the same way but they're sort of unrelated. The fact of the matter is Netscape does not equal Sun, does not equal IBM, does not equal Oracle, or these other cats. If they all had exactly the same middleware or the same...What should I call it, OS on an OS? That would be more competition than the fact that they're all a little bit different. Both Sun and Netscape are trying to get a position by riding on top of Windows and obsoleting it. And we still very much fear that as an approach.

Sun may be the guys who made Java; they're not necessarily the guys who are going to benefit from Java. If you actually asked me about the server side--who is going to make the Java middleware that really gives both Unix and NT Server the biggest run for their money--if I was handicapping, I'd probably say either Oracle or IBM or even Netscape more likely than Sun. IBM has more middleware at the server layer. If you asked me who is more likely to make a serious run with Windows at the client, I'd probably say it was Netscape taking advantage of Java than Sun taking advantage of Java. So even though Sun may own Java, it will be perturbations--I know I'm not supposed to say that in the Java world--but it will be Java plus some other things that probably put the most serious pressure on Windows: Java plus the Netscape browser, Java plus company Foo's middleware.