Executive Q&A
2003 Financial Analyst Meeting
July 24, 2003

CURT ANDERSON: Okay, we're going to get started with the Q&A. ... We've got a couple of the IR staff coming down the aisles here with mikes, so go ahead and just raise your hands when you've got a question and we'll call them out and try to get to as many as possible. ...

QUESTION: With your focus on integration and your success in the server market, could you give us your thoughts on getting into the enterprise applications market?

STEVE BALLMER: I think we've been pretty clear—I hope Doug was clear today—we're enthusiastic about what we're doing in Microsoft® Business Solutions. We've had a clear design goal to focus in on small companies on up through midmarkets and maybe what I might call smallish, smaller enterprises, and we think that that's the right design point for us. There is a higher-end design point that would require products that look a little different than the products we have today. But we choose not to focus in on that part of the market. And I think that works—partnership with SAP, partnership with Siebel, with PeopleSoft. Partnership with Oracle is probably a little harder, but we'll see what happens.

CURT ANDERSON: Another question? Down here. If you could raise your hand high so that these guys can see you and get a mike to you in time.

QUESTION: You've made a number of strategic investments in the past, but one of the areas that you recently indicated was particularly important to you, yet I didn't see anything in terms of investments, was the healthcare vertical. And so I was wondering if you can talk a little about that as well, as I've heard that Bill tends to say financial leverage begins with architectural consistency. So now that .NET has been around for a while and there are applications vendors beginning to provide somewhat more mature products to the marketplace, can you talk a little bit about some of the areas where you're seeing some early success on that platform with some of the ISVs?

STEVE BALLMER: I'm not sure I understood the second part of the question. Let me try the first, and if one of the other guys did they can take the second part.

The first part of the question: I think healthcare is an interesting market and the key for us is to make sure that we take our platform at the highest level that's appropriate and put it in the hands of good independent software vendors who can tailor it for market. We did form a healthcare vertical organization formally this year or the last few months on a go-forward basis. I think that will be an important part of our activity. We hope to have the kind of extensibility in our Microsoft Business Solutions product line that people can write healthcare-specific applications using that as a foundation. And certainly it's well-known that the healthcare part of the economy is one of the biggest and, I might add, one of the least automated. So over time it could well provide good opportunities. I don't see us doing specific vertical applications ourselves.

BILL GATES: One thing that's pretty hot in that area is the idea of portable tablet device; the large screen size and the mobility have strengthened the work we're doing with the ISVs where you take patient records and are able to move those around, as well as the requirement now for these digital patient records is forcing an upgrade to those systems. So, you know, I think we've got a broad set of design wins. I'd say .NET is particularly strong in that vertical.

QUESTION: (Off mike.)

BILL GATES: We can get you the information, but off the top of my head—

STEVE BALLMER: I would go to Microsoft.com, punch healthcare, and you'll see a full list. The fact of the matter is most of the ISVs in that space are much smaller. It's not like there's—I mean people say, "Well, what about, you know, SMS, the new SMS, what about Cerner?" Those guys are important; we have some pretty good traction. But most—you know, you take patient information handling: They're not really front and center in most of the scenarios Bill described.

CURT ANDERSON: Let's go here to Number Three, please.

QUESTION: A question about corporate adoption, the rate of adopting, particularly in Windows® on the client and Office. We're clearly in the middle of a big product road map, and yet the value that businesses realize depends on their willingness to accept your new scenarios and deploy them. How do you measure, particularly as corporations get on subscription agreements, how do you measure whether they're enjoying the fruits of all your instructive scenarios? And how do you measure the rate at which companies are taking you up on the new scenarios? Just looking around the room here, most of our big businesses seem to be a couple of revs back, if not more, even if we own the software, which we may under an enterprise. So I'm trying to keep it clear. How do you drive the adoption of the new scenarios and how do you measure how you're doing it, helping customers realize the value of the new scenarios that you're building?

STEVE BALLMER: One of the critical objectives we gave our field in the last six months is to really make sure they understand and drive deployment, not just purchase. And so we are literally currently in the process of asking, at least in our larger accounts where we tend to have the highest percentage of subscription agreements, that systematically detail what the deployment rate is or what the deployment situation is for various releases of the operating system. It's not statistical, because they won't let us instrument their networks at this stage. It's not 100 percent automated. But, you know, we're going to watch that, drive that. We put in place the business productivity advisers last year, about 400 folks, whose Number One job is to help show the kinds of new value that large accounts and smaller accounts can get out of deploying the new software. So we have a team focused in on it and we're building at least our own internal reporting measurements on that.

Just a small show of hands, because there's a lot of laptops, I noticed, open, which is a great thing: How many of you are running Windows XP on your laptops in front of you? And how many people are running Windows 98? Windows 2000?

Okay, so maybe what we have here is probably 40 to 60, something like that, which at least I like the deployment rate. The deployment rate in here is not perfect, but it's not that bad either. (Laughter.)

JOHN CONNORS:. But say for Bank of America, the account team's job is to know where you guys are at in terms of contracts, in terms of desktops covered, in terms of versions that are running. And it isn't automated and instrumented yet, but each account team, particularly for large accounts, they know where we're at

BILL GATES: Anybody who uses SMS, the SMS inventory function delivers that. And we're getting and we certainly have a strategic goal to get very high penetration of SMS.

Another thing that I found fascinating is that Office 2003, the latest version has, as we get that shipped, it's got built-in logging feature that we call SQM that tracks exactly how Office is being used. And so a corporation can roll that up and we give them tools to look at it and say, "Okay, what kind of Outlook® performance are they getting? How many applications are people building in these things?" And even they want to know that information is there. And then our BPA adviser steps in and can say, "Okay, your network, you ought to look at the reliability or you ought to look at some of these real-time things that your people are using less than other people in your vertical industry segment." So that ongoing relationship that's part of the systems assurance is something that we're driving value into.

CURT ANDERSON: Okay, let's go with Number One, please.

QUESTION: With regard to R&D, how do you guys determine what the right level is? And once you determine that, how do you go back and then determine what the return on that investment has been?

STEVE BALLMER: There are sort of two sides, so I'll give you both sides. The first thing we do is we have a sense of how many people we can hire. That's probably the Number One ceiling, frankly, because there's always plenty of other things people want to do.

Number Two, we're not irresponsible. We do take something of a view of what we can afford, although the things that tend to get caught tend to be on the sales and marketing side more than they tend to be on the R&D side. And then that sort of all gets rolled together with the business group and Bill and his staff, enthusiasm for new ideas, and particularly on acquisitions, where you can essentially get a lot of people quite quickly. Then we kind of munge it up to haul up and move forward.

There aren't many things that we think—that I can think of that we haven't funded where we actually thought we could execute well and there'd be an opportunity when we got through it. I don't know if you want to add.

BILL GATES: Well, I'd say you can say that success has given us a little bit of luxury in this area, that we go to the product groups and insist that they have a lot of incubation activity and we review the incubation activity, make sure that it's at quite a level—that they're doing risky things, some of which don't get past the incubation stage. The things that are particularly risky are done in the research environment, where we don't even have particular time frames of product shipment that the technology gets into.

Amazingly, I can actually explain exactly why we have every head count in MSR even somewhat better than in the product groups. That is, the product groups, what they're doing is the right thing, but we need to drive more efficiency into that. Whereas in the research activity we understand the efficiency equation very well, and we've gotten an immense amount out of that.

STEVE BALLMER: And, in fact, the largest, on a percent basis, I think, the largest grow-through in terms of R&D will actually be research this year.

CURT ANDERSON: Let's go to Number Two.

QUESTION: How do you position "Jupiter" into the ISV product road map and how are you positioning it sort of in conjunction with Windows Server™ and Visual Studio® .NET? And then, theoretically, how would that dovetail into "Longhorn" down the road?

STEVE BALLMER: I think the Jupiter concept—because it's not a product yet, it's a concept—I think the Jupiter concept is a concept that really is tailored at large enterprise customers more than individual ISVs. And I think as we continue to refine both our Windows Server plans as well as our Jupiter plans, we may find that some of the technologies that we pioneer in Jupiter will become more a part of the base Windows Server fabric in order to be more of a real platform for the independent software vendor for whom it's hard to say, "Yeah, I'll take another higher-priced item and make it a fundamental part of my product definition."

You know, there's probably—Bill talked about workflow, I think, a little bit earlier today. There are elements at sort of the basic level of workflow that probably ought to be available to ISVs, not just available in the higher-end SKU, as an example.

BILL GATES: We're able to take subset runtimes, like we did the SQL Server™, and place them very aggressively for ISV usage. And so you'll see the key parts of the things in Jupiter where you will find a way to make it attractive to ISVs, particularly in the workflow engine orchestration space because that's in the process of moving to be part of the platform itself.

CURT ANDERSON: Number Three, please.

QUESTION: Could you tell me how you're going to sell more Xboxes in Japan? And second, can you give me more of an idea what sort of a direction you want to go in the antivirus marketplace, please?

STEVE BALLMER: I guess I'll do Xbox® in Japan. (Laughter.) We think there is—right now what we have in Japan is a small niche market—happy customers but small niche market. And we're going to try to build off that base of customers and their profile. A lot of them also would own a PlayStation, for example. They like some of the unique games that are on our environment. We're going to push on that.

I'm not sure, you know, in the next few years we will characterize this as being a very large player in the Japanese market, but certainly we will be larger than we are today.

We did bring on board a new vice president to oversee the international operations for Xbox, a fellow named Peter Moore, who we're very excited about. I think Robbie may have mentioned him in his conversation earlier today. He's taking care of not only Japan but also Europe, and he's doing some things to enhance our developer relations and the team in Japan. And I think we understand that that'll be a market that is a tough one where we're going to have to work on it for a number of years.

Second part was antivirus. Maybe I'll let you do that one, if you don't mind.

BILL GATES: Okay. We've offered antivirus capabilities to the customers under certain OEM arrangements with vendors in the past. MSN® has done that, as have one or two other groups. We also now have an antivirus group that's part of Microsoft.

And so, as we're refining our Software Assurance offerings and consumer subscription offerings, and as we're enhancing that technology to a new generation that goes beyond how people have thought about it, we will find a way to get that to our customers, primarily through subscription-type offerings.

STEVE BALLMER: And I think you ought to think what we care about is that our customers have the best security experience in the world. We've decided we need more capability and we will grow kind of—if you will—the footprint of our offer, but we know it will continue to need to be complemented. We'll want to see it complemented by offers from others. So we're in pretty good dialogue with guys like Symantec and NAI, et cetera.

CURT ANDERSON: Number One over here, please.

QUESTION: Can you elaborate on how Microsoft uses total cost of ownership of less than Linux and how you position that in competitive situations and with customers in getting them to see those cost arguments?

STEVE BALLMER: Well, I did the brief form and I probably won't do the long width version, but if you step back and you say, "Look, let's see what the independent research organizations see as cost," there's not only the cost of acquisition, there's the cost of deployment, there's the cost of ongoing operation, there's the cost of other value-added software, transaction software, middleware, blah blah blah, that you want to add on top of this that comes fully included in the basic price of the Windows product.

If you take a look at that, you get the kind of results—and you can go through with customers quantitatively, "Okay, here is the acquisition cost differential, where Windows Server is, say, 500 bucks roughly, but what about this, what about this, what about this, what about this, what about this, what about this, what about this?" And that's how you get to the kind of delta that we see in terms of total cost of ownership both at the client and the server and you just walk customers through that stuff systematically.

Sometimes that's very effective and sometimes it still can be problematic if the customer sees a very easy migration.

The place where we have the toughest time is when a customer thinks it's easier to migrate an application they already have on UNIX to Linux than our stuff. Now, we've had good success migrating some customers, but there are still cases where either the reality is or the customer's belief is it's easier to migrate to Linux, and we continue to work from an R&D perspective on our technology to facilitate those things, and we continue to work on our ability to articulate the cost of migration.

CURT ANDERSON: Let's go to Number Two here, please.

QUESTION: Yes, two questions: One is how long will it take to get all of the Business Solutions products on a single platform, .NET, and also how important is that to gain share in that fragmented market?

Separately, on the subject of India, how much do you plan to use that going forward and what can it be used for and what can't it be used for? Thanks.

BILL GATES: The companies that we acquired to build up Microsoft Business Solutions were companies that were extremely good users of Microsoft technology, so, for example, the way that Great Plains® used our platform, the way Navision® used our platform, and so they already have extensibility capabilities that use .NET. They're already exposing Web services.

There's a fairly big generational upgrade to our applications platform that comes in the Longhorn time frame and that goes a step further, in fact, taking advantage of the new things that are being developed in parallel with it. But we already view those as a very strong showcase of the .NET runtime.

JOHN CONNORS: Yeah, in terms of India we've had a presence there for some time with a sales subsidiary and a research and development site. We also have there some call center product support and internal IT application work that are Microsoft employees.

The majority of the work that is in India was done previously by vendors, but I think India, China, and Eastern Europe have a large number of very talented technology workers, as well as computer scientists, and you would expect us, like others as being global companies, to have a larger presence there. And I think all companies, including us, benefit as those economies grow and more people reach middle class. So R&D, there is some. There is some internal IT development and call center and support, and you'd expect to see that grow.

STEVE BALLMER: Yeah, the R&D centers, I think there are 150, something like that, people now. And we can see 100-150 more, roughly. We can see maybe doubling, tripling that in the next few years. But we're not sort of envisioning that that gets to be 10,000 people over the course of the next several years. I think we will be able to more—we will make more use of resources in India, both our own sites as well as vendors, in the areas and other areas that John highlighted than we will see in R&D. And we'll try some other experimentation and see where that goes. We have a great research facility in China, a great research facility in China, and that we do continue to grow as we grow Microsoft Research.

QUESTION: Two questions. On the porting topic, John, could you talk about how much money is being invested in that SFU product or related products?

And then, secondly, Steve or Bill, could you talk about the applications market, particularly the high end. Orlando was quoted a couple of weeks back to the effect of, we're coming out market, and we're on a collision course, which seems to run counter to some of the comments you've had in the past. Could you give us a sense of what your intentions are there? Thanks.

JOHN CONNORS: Somebody else can take the SFU question, because I can't remember what that's an acronym for.

BILL GATES: SFU is our UNIX emulation capability, and that's exactly what Steve was talking about where we can say to a corporation that's got, say, Solaris systems, they move down to Windows, and we don't need to do much work with your applications because they have all of the shell capabilities, utilities, and API capabilities that you'd have up on the Solaris system. And we have a lot of great design wins—that's another thing that's well documented on the Web site. The team that does the bulk of that work is in India, and I would say there's probably 70 or 80 people there. We also, as part of our license with the Santa Cruz operation, not only have the intellectual property rights to make that emulation layer clear and free, and part of the indemnity that we provide to our Windows customers, we've also got the SCO source code to charge up that effort and go even faster in saying that, yes, we provide UNIX capabilities, but we also provide Windows, which has a lot of things better than that.

STEVE BALLMER: In terms of the second question you asked, everybody wants some black-and-white answer. We have a design center that starts with smaller and midsize companies that we're going to sell our products to as much as big companies can productively use. SAP, as a design center, gets large enterprises, and they're going to sell their product line to companies that can productively use it as small as they get. There is an overlap, there is an overlap space. I think that overlap space represents a very small percentage of their revenue, and not a very big percentage of our current revenue. And that market is very fragmented, where we've seen more of J.D. Edwards, and a lot of other people. So it will be an area of some overlap, but the design center for our development is different than their design center, and that's why the two companies continue to have a very, very productive partnership even though we know they're coming after us in the small and midmarket, and they know we're coming up a little bit, but still very productive type of partnership.

QUESTION: (Off mike.)

BILL GATES: Linux is UNIX, and it's a form of UNIX. It's not like anybody invented a new way of doing an operating system. It's like Free BSD was, and that existed even 10 years ago. The open source approach is valuable for certain types of development. We've always seen software coming out of the universities that prototype those things, and it's part of the ecosystem of our industry that that was there, and used, and people would, if the licensing model allowed it, if it was actually an open license, which the open source license is not open because you can't take it and ever use it in a job-creating activity.

Anyway, people did take those things and there was kind of an equilibrium between what came out of that and what was done on a commercial basis. I don't think that's going away. There is a clash of innovation having to do with management, security, natural user interface that goes beyond what UNIX has delivered, that we're driving forward. It's like designing a 747 or a moon shot, where there's a lot of integrated activities, the design, the testing, the pieces all come together giving a version of Office that exploits those things. And our innovation leads to remind people that what they expect out of an operating system isn't standing still. And so there's an onus on us that if it's still 1970s UNIX that a lot of people expect, then we'll share more of the space with that kind of approach than if we're revolutionizing and getting a whole new level of expectation. We were able to do that in things like word processing and spreadsheets. We're in the process of doing it in categories like database and collaboration. And, so over the next four or five years, that's why we've taken the road of pouring on the R&D, taking advantage of the hiring environment we've got today, and driving this effort.

In the next four or five years, people will understand some of the IP questions that get raised about those systems, and is there some sort of patent pool that people are paying royalties into, how that all plays out so that customers don't feel like they've got this open-ended liability for something that comes without that indemnification. I don't have an exact prediction for that, but that's something that equates some notion of rights and friction into that side of the system.

STEVE BALLMER: One thing I would like to comment about, which is: We like lots of customers. We're going to invest to have lots of customers. And, you know, Bill said, will we have a bigger piece, a smaller piece, we're going to prioritize to having as big a piece as we can have based upon what we offer the customers. And there may be occasionally some friction between that and revenue and profit enhancement. Will we include more value instead of trying to price something separately from time to time, to have sort of a more valuable offer as opposed to doing something that might look like it had greater revenue optimization short-term? Expect that kind of decision process around here. People said, hey, look, are you really—we're doing some things to try to really reach out, particularly in poorer countries, to make sure that our products are getting used by folks in education, and you'll hear more about that. Well, that can't be done at the same kind of prices that people might expect in other environments. I'm trying to emphasize, we're prioritizing having lots of customers, and when I talked about balanced goals, that means tradeoffs sometimes with short-term business growth.

QUESTION: Right now it seems like there's a land grab for sort of the VAR and reseller channel—and also some of the smaller ISVs—as some of the larger infrastructure players move down. You've traditionally had a very large share of that market, how do you plan to protect that, particularly when in the ISV side you're competing increasingly with your ISV customers? And my second question, can you give us a little bit more detail on your Dynamic Systems Initiative, and what that could mean down the road?

STEVE BALLMER: Maybe I'll take the first, and I'll let Bill have the second.

Yes, we're investing to try to drive—we have a good, as you say, a good, strong position amongst the small and medium—the VARS who serve small and medium businesses, that has been a stronghold of the ISVs that serve those companies. And there's really sort of three things we're doing. One, John talked about a set of investments we're making to dial up our activity directed at small and medium customers; most of that, frankly, is going into enhancements to the programs that we have in place for our partners. So that's Number One.

Number Two, particularly for the software developers amongst that community, we are building a set of new programs and support mechanisms that we think will be very important, not just around Windows and SQL Server, but also around the Microsoft Business Solutions product set itself, because, as Bill said, that stuff has been built around .NET, it is extensible, and we think a lot of those ISVs, it's a win-win, instead of looking at us as a new competitive threat, looking at us as a source of opportunity to do vertical applications, et cetera, that come on top, and not reinvent the wheel as much down at the base. That would be probably the second thing I'd highlight.

The third thing is, we're trying to crank up our number of customer adds with Microsoft Business Solutions, which is in some senses the more success we have, the more it becomes a train, a piece of infrastructure that other ISVs absolutely want to add on value on top of.

BILL GATES: On the Dynamic Systems Initiative, this is the idea that instead of humans having to set up systems and track which systems are more loaded, or what problems there are with the systems, that instead you would have a piece of software that's operating across all the different systems, whether it's in the data center, across the desktop, across all the data centers in a corporation. That software would understand exactly what your goals are, in terms of getting tasks run, and providing a certain responsiveness. And therefore, allocate the hardware resources in a very dynamic way.

This initiative is a very exciting thing. Our research group started this work about four years ago in an initiative they called Billions of Interacting Gizmos. If you have billions of gizmos, you could never buy operators to manage them, so let's design this so that it can work without operators at all. And we've actually shipped the base platform for that, with Windows Server 2003, and now we have some add-on management software that will come in and use those hooks to do exactly that kind of dynamic provisioning.

The work here involves modeling what the hardware is, modeling what the applications are. It's a huge piece of R&D, and we are making really quite rapid progress on it. It's a pure software problem; it's not a hardware problem. Interestingly enough, some of the other companies, more hardware-oriented companies, like IBM talks about it as autonomous computing, or Sun. We feel that the road map we've got, the way we've got it integrated into the tools, we'll actually make that vision a reality far more rapidly than anybody else can.

QUESTION: I just wanted to follow up on an issue you touched on with Scott's question, and that's the IP issue on Linux. I'd be curious whether you're seeing customers where this is actually starting to sway decisions, the uncertainty right now, and as people begin to focus on this as an issue. Thinking on patents, where Microsoft is protected to some degree by the cross-licensing, I'm not sure Linux has that kind of protection of cross-licensing and patents. Is this just the beginning, do you think, for Linux, of these kinds of issues? And it's not clear to me how those get resolved.

STEVE BALLMER: I'm going to make a one-sentence comment about the customer reaction, and let Bill really answer the rest of the question. I think we've seen—the average sort of person on the IT floor of any of our customers absolutely knows there's issues around this. They don't understand them, but they know they're issues. Primarily because of the SCO-IBM lawsuit. So I'd say that. I think in terms of really understanding anything about the details or the risks, you now have some analyst groups that are sort of explaining their view of what this might mean, not just SCO or IBM, but more broadly the patent and IP issue. But, it's still not a very well-developed understanding that I think I see in our customers. Maybe it makes sense for Bill to spend a minute and give an explanation of how we see things.

BILL GATES: A little bit of background on this. In the technology field the larger companies invest very substantial amounts, not only in R&D, but also in getting patents that reflect that R&D. So a company like Microsoft, last year we applied for 1,500 patents, and that's a number that's been going up at a pretty steep ramp. The number you actually see disclosed lags that by an increasing amount, it's about three to three-and-a-half years now, but it will probably go up to more like four to four-and-a-half years, because of the application process in the U.S. Patent Office. And so there's a huge investment in the R&D, and in getting the patents.

You might say, "Well, why hasn't the technology industry devolved into thousands of patent lawsuits, where Microsoft sues IBM or vice versa, relative to this innovation?" The companies that are really doing R&D and are serious about it have by and large—there are exceptions—by and large entered into types of cross-license agreements. And so, for example, IBM and Microsoft did a cross-license agreement I think, eight, nine, 10 years ago—a long time ago, when we were small. And so that works well. Linux is not covered by most of these patent cross-licenses. So, for example, when we license somebody we cover their commercial products, not the noncommercial products. And so by and large, the main place you've seen patent lawsuits has been companies that are not participants in the industry, but rather sort of a lone inventor-type activity. And because there's so much priority, often those things end up not creating a liability for the company.

Here's where you have a product that's not—it doesn't have an R&D contribution, it's not participating in these cross-licenses, given the breadth of functionality, you'd expect a very high percentage of companies in the IT business would have patents that would read on the functionality in that product. And so it's a new thing. One thing to understand about the GPL is that you can't just partially license—somebody can't go and just license IBM Linux or Red Hat Linux, the way the GPL works, if you license anything at all you have to license it all. And that means, essentially, getting nothing in return. That's the only way that it can be covered is basically for you to decide just to give that piece away.

QUESTION: The GPL being the—

BILL GATES: The unusual license that is the one that has the word open, because it's not open, is called the GPL. So anyway, the whole IP thing is just beginning to get attention, because it is a—it's not a scenario that existed in the past.

QUESTION: A couple of questions: How concerned are you that some of your intellectual property may, in fact, end up in some open source products, and how close are you looking at that? And then as a follow-up, you are one of the few out there who has access, obviously, to SCO's source code, you can obviously look and see what the Linux source code looks like. In terms of your investigations, could you talk a little bit about what you're seeing, in terms of code that looks like it was directly taken from SCO?

And then as a totally separate question to John, with respect to your comments on decline sequentially in deferred revenue next quarter, if my recollection serves me correctly, I thought in the September quarter a year ago about half your increase in deferred revenue was associated with EAs, about half associated with the older licensing. I guess I'm a little confused as to why you would expect your deferred revenue line to decline so significantly sequentially?

JOHN CONNORS: How do you guys want to answer?

BILL GATES: I'll take the first part.

STEVE BALLMER: I'll take the second.

JOHN CONNORS: I'll take the third.

BILL GATES: Certainly there's no question that, particularly in some of the more cloning-type activities, intellectual property from many, many companies, including Microsoft, is being used in open source software. It's pretty much when people clone things that often becomes unavoidable. The SCO suit, is really—you had a comment on that subject, that's largely related, or there are aspects of it that are unique to them, because they relate to trademark and copyright.

STEVE BALLMER: We haven't asked our people to go study through the source code and comment on some lawsuit. We're using the software for the purposes that Bill talked about earlier.

JOHN CONNORS: We do anticipate the unearned balance to tick down in the September quarter from June, in part because that's just the normal seasonality. And we had quite a strong Q4 in the enterprise space, and the government space. A year ago we had the abnormal seasonality in unearned, largely as a function of upgraded advantage purchasing ahead of the Licensing 6.0 transition, not EA. So we don't have any similar phenomenon this year. So the year-over-year growth rates are really not comparable due to the Licensing 6.0 transition.

CURT ANDERSON: Let's go with Number Two up top there, please.

QUESTION: Yeah. Bill talked a little bit about the silos of information as being a real issue for business users, and we've been a lot of that. Are you thinking that you're going to try to help solve this through driving XML adoption, or are you actually going to try to create products that attach and extract the information from other application software products?

BILL GATES: We'll do both of those things. The most popular business intelligence viewing tool is Excel, and we've done some very exciting things in Office 2003 to connect Excel to Web services. In our view that's just the start of making it possible for knowledge workers and enterprise to really know what's going on, which is an unsolved problem: to see sales trends, to budget very easily, to analyze customer feedback data.

And so both the Office group and our server platform group, particularly the SQL people, are investing very heavily in tools that extract data, model the data in a simpler way that the user understands, but yet make it so it's mappable back to the very complex tables in the ERP type software—and so sit and look at sales in a normal way, but if you want to change a forecast or something you have write through into that infrastructure.

So it's a very interesting problem for us. BI has been a sort of modest volume, high-price-type market. We hope to as we have with most other knowledge worker things turn it into a very broad, high-volume market with very modest costs per knowledge worker.

QUESTION: (Off mike)—Oracle, PeopleSoft situation and how it would affect the industry.

STEVE BALLMER: Just our view? (Laughter.) In a sense, it seems cosmic to us, as to you. In a sense, you know, if you actually ask what do we change about our strategy, it doesn't change nearly as much as you might think. We still have to work well with Siebel. We're still going to have the challenge of getting Oracle to do anything with SQL Server. That ain't going to change. PeopleSoft, we don't have that high penetration. It's not like sort of an immediate big thing at risk.

We're trying to build our penetration to SQL Servers, under PeopleSoft, but we don't have that. And, you know, we're going to continue to work hard to have relationships with SAP and the variety of vertical—at the enterprise base. So it doesn't feel like it changes our strategy all that dramatically, but it's certainly something that, you know, everybody in the industry is talking about, thinking about, services, companies, you know, software companies, hardware companies, because everybody's not sure what other dislocation might be associated with it.

CURT ANDERSON: Group Number Three, please.

QUESTION: It seems that just as software has a synergy with Moore's Law, content has synergy with generational increases in bandwidth. In that, it seems that Web services is somewhat analogous to object-oriented programming in allowing you to keep up with providing content for bandwidth. With the success of .NET largely dependent on increases in bandwidth, what do you see as possible solutions to the last-mile problem?

BILL GATES: I assume that—there is not a bandwidth problem connecting businesses in the developed world. And also, Web services, a lot of what flows across Web services is transaction-type information. And that is so small. When you're flowing to a bank, when you're flowing to a retailer, when you have a supply chain, the actual number of bits that are flowing there is so modest it's unbelievable.

And even a dial-up connection is enough for consumer-type scenarios for buying and selling, and those things. So the main place you get into a bandwidth problem is preventing multimedia content and making it an always-on connection that's convenient to just go and use—space, taking up the phone or the newspaper and doing commerce or information browsing a different way.

So we're really big fans of all the new approaches that would drive broadband to be more popular. Getting the DSLs, our guys to do a better job and have a rich offering, work with some of the cable modem companies to have better offerings. But we don't see that as a real bottleneck in Web services adoption. It's actually more of the security infrastructure, that billing infrastructure, getting people used to these development tools. That's the piece that clearly is happening, but it takes time, insofar as mainly the pioneering customers that are now having a positive impact on the rest of the customer base.

STEVE BALLMER: Broadband, I mean, let's face it, has been exploding more in some other countries than in this one. But even in this country, the number of broadband users is growing at a very healthy rate, really everywhere in the world. You look at places like Korea, Japan, what's happening there, Germany. But even the United States is growing quite nicely.

CURT ANDERSON: Number One, please.

QUESTION: Yes, a question on the federal government budget. We are hearing there is some consolidation in the number of purchases, down to 20 from 250. And some companies have been able to capitalize and federalize a budget, and some companies have been unable to capitalize on the new purchasing, the way IT budgets are being purchased from the federal government.

I was wondering, what are you seeing in IT, federalizing spending budgets? Do you think there will be some budget selection at the end of the October time frame?

STEVE BALLMER: Well, I think the U.S. government's a very hard organization to manage, and the appropriate officials are working very hard to try to figure out, not just on IT, but in a variety of senses where there might be ways to get expenses under control. Certainly, we're a big supplier to the U.S. federal government. We have been for years and years, probably since Windows—what, 2?—when the U.S. Air Force was an early adopter of the Windows product.

We've continued to be a large supplier. And we certainly see the government trying to be smart and appropriate in its purchasing, and we're working very well with the U.S. government on that. I can't tell you what the total—I mean, there are people in the U.S. government who will tell you what the total IT budget is, but you can get different people to give you different numbers, frankly.

So I'm not sure I'm the right expert to tell you what the real stat is.

JOHN CONNORS: And I think our business has been quite good in the U.S. and overseas. And I think, in general, we've got a very good federal model, both from an accounting perspective and a technical support perspective. Because we had this question previously, a couple of months ago, and our team felt that we were in very good shape with whatever ultimately ends up being the legislation that Congress passes for consolidating purchasing. We'll respond accordingly.

STEVE BALLMER: And, you know, we'll work with the customers if there issues and we need to address them commercially, we'll address them.

BILL GATES: We did just announce our biggest customer contract ever with—

JOHN CONNORS: The U.S. Army.

STEVE BALLMER: Great. Yeah, biggest contract we've ever done.

CURT ANDERSON: This gentleman down here had a quick question.

QUESTION: You mentioned earlier that there is a new layer of CFOs within all of the organizations, and it's evident that their charge is not going to be for cost-cutting throughout the organization. Could you elaborate on what, in fact, you do expect out of these executives?

STEVE BALLMER: I want—I'll speak, and then John can be more specific—but from my perspective, I want somebody who's the voice of finance thinking through smart business deals, good expenditures with smart use of resources, making sure that the rest of the leadership team is really thinking about where revenue's coming from, where costs were going, cutting costs when costs need to be cut.

But, you know, we didn't say, oh, we need cost cutters; let's go put cost cutters. But I expect those people to think intelligently about those businesses financially, in about the same way John thinks intelligently about our business overall. And I have to say, John's more of a conscience on making sure we really watch our overall costs and are thinking smart about—you never want to waste money, you want to spend money on the important things.

The size of the cost pool is not as important as making sure that every dollar is working hard for the customers, for the shareholders, for somebody. And certainly, that would be part of those divisional CFOs. But I want every one of our business leaders to have a partner, the way I have a partner in John, thinking through the finances of the business.

JOHN CONNORS: Yeah, and I think if you think about how the company is large today, and we have a number of different, diverse businesses, these guys have to—they've got to build great teams or grow the teams they have. They've got to have a great controller. They've got to ensure we have a great control environment. They've got to be sure we're doing smart deals, not only when we buy them, but then that we integrate them well.

They've got to be very, very effective at planning and strategy. They've got to challenge the business people. And we do expect them to be very good at either having on staff or themselves, how do you reengineer the way we work? How do you reengineer the way we think? And, in general, be very good finance people, but great business people. And they should be growing a whole generation of great business finance people below them.

So we've got a broad mandate for them, and I think the folks are up to it.

CURT ANDERSON: Number One, please.

QUESTION: Thank you. I have two questions on timing, and the first is with the introduction of the "Yukon" database. When will it be introduced? And is its introduction independent of your Longhorn initiatives, or simultaneous with Longhorn? And then, the second question has to deal with legal issues. Can you provide us a road map or a timeline about what the next big events are for your legal issues? And then, what is management's commitment to getting these legal woes behind it?

BILL GATES: Okay. Yukon is a code name for the next generation of SQL Server, and that product went into beta recently. And we'd expect to have it out in the marketplace by the end of next year, that is calendar 2004. Some of the Yukon technology, in fact, is contributing to the unified file system called Windows File System, which the acronym is WinFS. So Yukon predates Longhorn, which there's no formal schedule for at this stage. But it comes after. It'd be certainly after Yukon.

In terms of legal, you know, we're a big company and, you know, you pick an area, whether it's people who are pirating our software or counterfeiting our software or, you know, patent disputes that arise from the normal courts, or things like that. We have a lot of different things. At this point, you know, we've got a lawsuit with Sun that will probably go to trial in 2005. We've got a discussion with the European Union about any concerns they have and how we can get them comfortable with all the ways we offer our products—

STEVE BALLMER: Which is their timing.

BILL GATES: Yeah. It's totally up to them. We don't control that timing. We have some state-level, so-called overcharge lawsuits that we've been able to settle some, and a number that we have not settled, so they have different legal status. You know, there's a bunch of other lawsuits, as well.

STEVE BALLMER: I would say, with respect to those, our basic agenda would be to settle matters that can be sensibly settled, you know, whenever possible, particularly with governmental authorities. I mean, let's break into the two categories. There's sort of governmental authorities, and there's other plaintiffs. And we're trying to work very constructively on issues.

We will go through a litigation process, I guess, when we need to, as evidenced by the long process we went through that eventually led to the consent decree. But I think our desire would be to try to amicably settle those matters. In terms of private lawsuits, you know, we have to do the thing that's right for our shareholders, and sometimes that will be to settle, and sometimes that will be to go through the litigation process.

Everything is probably settleable, but not everything on the private side is settleable in some way that is fair to where we are and what's right for our shareholders. Some of these things have no merit. They ought to be litigated. We're prepared to litigate when we need to.

BILL GATES: And that's fairly important in the sense that if you settled everything, and management said, "Oh, we settle everything," then the only thing you could guarantee is that the supply of lawsuits would rise.

CURT ANDERSON: Number Three up there, please.

QUESTION: A question on migration of value as a platform business. And you talked, I guess, a couple times in the presentations and Q&A about the notion of Linux and software as a commodity versus Windows as a source of innovation with security and management and a new UI integrated experience. But isn't there a big percentage of activity in, like Windows, or under any platform R&D, like device drivers, bug fixes, where those activities can consume as much as 50 to 70 percent of the effort, that those are more maintenance activities?

And couldn't a community approach apply to that and an innovation approach apply to the value add, the new stuff on top, where you or someone like an IBM could add value on top of what the community maintains?

BILL GATES: Well, first of all, the last thing somebody at home at night having fun writing software wants to do is maintenance. I mean, if there is anything that the free software model is not good at it's maintenance. The idea that you take a system that's not working, you diagnose it, you dig into it, you have all these systems to gather the information, if a customer needs a fix within an hour you get him a fix within an hour; that kind of work you've got to have commercial relationships where you guarantee those things.

When we fix things we run a massive number of tests. Writing those tests is not the kind of thing that people at home at night get a kick out of saying, you know, "I'll just be a test writer."

So maintenance is not a huge part of what we do on Windows. The device drivers are largely written by third parties.

STEVE BALLMER: It depends by community.

BILL GATES: Yeah, by a community of people. HP writes printer drivers. People who do the peripherals write most of the peripheral drivers. NVIDIA writes the graphics driver software. So we're not involved in most of that activity.

Most of what we're doing are innovations in the operating system itself, but we have to guarantee that we have teams that are available 24 hours a day to fix things and that we are watching in this way that is not exciting but it's commercial important, we're watching everything that's going on with our system, you know, taking every one of those support calls, taking visits to the Web site, taking the feedback reports we get called Watson and digging into those things.

And the responsiveness people get out of that, you know, there's all sorts of statistics to show how well the commercial model does at that.

CURT ANDERSON: We'll take our last question. Number Two, please.

QUESTION: On the issue of cash balances—first, thank you for the additional disclosure on it—my question on that is there's been a debate about what to do with cash for some time. And during that period, the cash balances on your balance sheet just continued to grow.

How should we interpret that? And if we interpret that that you don't feel like you've had enough cash, is there a level that you feel is the right amount of cash? Is there a point where you would start returning everything from that point forward? (Laughter.)

STEVE BALLMER: I'm going to take this one instead of John—(laughter)—just to give a perspective. We've got to be very clear that we think there are some risks associated with some things that exist on the legal front and what they might imply and what they might bring, and I simply think it would be imprudent for us to make some kind of, you know, significant change to the strategy on our cash until we resolve those matters. There's no reason to do so. It's not going to enhance shareholder value to make those decisions.

The map on this stuff—you guys are all the experts, but Bill and I, we've spent a lot of time just looking at a basic map from John that says, "Look, if you buy back your stock and you keep the cash, once your stock is undervalued, do you have an improved shareholder value?" Everybody can have their own point of view on what undervalued is. And, you know, we show big buyback numbers. And I don't know, probably somebody's probably cheering and I'm sitting here looking and saying, "Gosh, we bought all that stock at 43?" Some year we had a bunch of stock we bought at 43. And, you know, that didn't create shareholder value.

The truth is, the way the math works, keeping a dollar in the company versus buying back if you're buying back whatever the long-term price is, those are economically equivalent. A dividend is economically equivalent. All these things are about giving flexibility to the shareholders in terms of how they deploy the cash.

And so, if it takes another n months—and I don't have to say what n is because I don't know—but if it takes another n months to resolve these issues, what we're trying to tell you is it would be imprudent to change something in our cash strategy now. We're not in any way reducing shareholder value during that period of time by our actions.

John tried to be quite clear that we think we're managing your—that is, all of our shareholders—cash responsibly. And we talked about the risky part of the portfolio, which is the so-called strategic part, which was—I forget—nine billion out of the total of 67 billion that John talked about. We talked about the kind of—you know, you wouldn't say it's the highest-risk fund that we manage in terms of our Treasury portfolio but, you know, it's a risk, it's a well-managed risk-reward relationship. No matter what we do, we're going to still do some of these strategic deals to take additional risk, no matter what our cash strategy is.

So what I tell everybody is we think it would be imprudent for the company to be hasty. There's no reason to do that. We're trying to be responsible during this period of time. And, yes, there will be a time at which, through one form or another, buyback dividends, some combination, we will give greater flexibility to our shareholders in terms of how they deploy their cash.

But, you know, John said, "Oh, everybody's going to expect an answer at this meeting." The money is being well taken care of, and we're going to get through some issues and understand really where we are. It's the only prudent thing. Certainly I feel that way not only as the CEO, but as a shareholder perhaps more importantly. And our board certainly agrees on that. Bill agrees on that.

CURT ANDERSON: Great. Well, thanks, Bill, Steve, John. (Applause.)

END

Due to the varying sound quality and subject matter of tapes, the information in this transcript may contain inaccuracies.