On March 22, four days after returning from negotiations with the European Union in Brussels, and two days before the EU's ruling against Microsoft, CEO Steve Ballmer sat down for an interview with InformationWeek editors at Microsoft's Redmond, Wash., campus. Following are some excerpts, which have been lightly edited for clarity and brevity.

The European Union:
InformationWeek: Can you tell us what the sticking point was over in Europe?

Ballmer: We and the commission agreed that we were able to, in our settlement proposals, substantively resolve the concrete issues that the commission had, but the commission was interested in getting something more formulaic in terms of guidance for the future. In the U.S., that's not the framework. In the framework here, there's not a formula that makes perfect sense. What was agreed to in the Consent Decree was a set of disclosures, obligations, and responsibilities, in addition to some other things, to try to ensure that there's both innovation and competition. And there was a set of guidelines articulated by the U.S. Court of Appeals in terms of the reasonableness of additional capabilities that we innovate and integrate into Windows, the reasonableness that the value they bring to some customers outweighs in some sense whatever impact they might have on competitors. We tried to speak to some of the same points in our settlement proposals, but the commissioner seemed to want something more concrete, more formulaic, and we weren't able to do that, even though what we offered in settlement goes well beyond what they'll ask for in remedy this week.

InformationWeek: Does this boil down to the question that's been raised before about what gets included in the operating system and Microsoft's ability to innovate?

Ballmer: In a sense, yes, and in a sense, no. There is now a standard of reason and a standard of contract in place in the United States. The question is, will the standard be different outside the United States than inside the United States? We went through a long and arduous process in the United States, including litigation, hearing from the District Court, the Court of Appeals, the consent decree, the modifications to the consent decree to have most of the states participate. So there is a framework in place in the United States that answers those questions, and we think it's quite an adequate framework. We had hoped to settle this thing with just the European Commission, so we didn't have to revisit some of the issues that have already been put to bed. Looks like we will be back in front of what's called the Court of First Instance, though, on these issues because we have made clear that we will appeal the commission's decision. [The Court of First Instance is an EU court that hears employment disputes and other matters.]

Vertical-Industry Strategy:
InformationWeek: We'd like to talk about Microsoft's strategy as it relates to different vertical industries. How would you rate Microsoft's ability to be a provider of industry-specific solutions? And how is that changing?

Ballmer: For small and medium-sized businesses we're trying to provide more complete applications capability through the work that we're doing in Axapta, Great Plains, and Navision. For that class of customer, where we have a full applications set, which doesn't include the General Motors of the world, we're trying to build up a horizontal base of capabilities that's easily customized for vertical solutions as well as a set of partners who are vertical in their nature. So we have great manufacturing software, we want people to build out the manufacturing software for clothing manufacturers. This is for smaller and medium-sized enterprises, as opposed to the largest of the large. And I think we have a very good base of partners and a good presence at that level.

For the largest set of enterprises, or for enterprises which for whatever reasons are not candidates to become customers through our Axapta, Great Plains, and Navision product lines, what we're really trying to do is reach out to partners to build out vertical capability. We aren't ourselves trying to be the provider of automotive solutions or oil and gas solutions. So the question is, in a sense, how well do I think we're doing with independent software vendors? And the answer there is, by and large, we're doing quite well, but always with opportunity to improve. I can certainly say we're in better shape now than we would have been a couple of years ago. We're more focused in on the verticals. We're more focused in on the ISVs and system integrators in those verticals. .Net has matured as a platform, which makes it more interesting to people. We have more of a marketing engine that knows how to speak to those customers, which is important to the partners that we're trying to address. So I think we've made a huge amount of progress on that dimension.

At the end of the day, we don't provide the vertical capabilities. I'll contrast this first with SAP. There are verticals where SAP has, for large enterprises, deep vertical capability. The process-manufacturing industry is probably the most concrete for SAP. If you look at chemical or oil and gas or some of these other so-called process manufacturing as opposed to discrete manufacturing, they've got a depth of verticalized expertise. We want to have a great partnership with SAP, but the vertical knowledge is really embodied in their applications. IBM's on some weird strategy, in my opinion. Weird in the sense they clearly don't have the vertical capability, but they're trying to suggest that they provide vertical capability. They do in their services, of course, but not in their software. So I think they're neither fish nor fowl. They're not a pure friend to the other ISVs and software integrators, nor are they really themselves trying to be a provider of software that solves these problems. So I think we've made great progress, we have great capability today that we didn't have a couple of years ago, yet I believe there's a whole lot more that we can and need to be doing.

The other thing I'm all hepped up on these days is what I might call the horizontal verticals. You say, what the heck does that mean? Take finance people in companies. A lot of finance people are heavy users of Excel. A lot finance people use SAP and Hyperion. But finance is a function that I don't think has gotten the same kind of verticalization that industries have gotten. Marketing, you could say, 'Isn't that PeopleSoft's job [or] SAP's job?' In a sense, yes, and in a sense, no. Those are more about line-of-business processes, as opposed to: An individual finance person, what would they want? An individual salesperson, what would they want? Individual salespeople mostly see the CRM systems as, shall I say, tools of oppression? Sales management loves them but salespeople don't tend to love them. They're there for somebody other than the front-line seller. And I do think there's going to be more work in a vertical sense working with these functional heads than we've had in the past. It's a different kind of verticalization, but I think it's going to be equally important.

InformationWeek: Would you address that through Microsoft products, or through ISVs?

Ballmer: A little bit of both. We're really taking a look at that now as a key next issue. Longhorn:
InformationWeek: We'd like to talk to you about Microsoft's value proposition. Is there a change in thinking going on that relates to the way technology is delivered to your business customers in terms of packaging or licensing? Longhorn looks like it's going to be a couple of years out and the product teams are talking about updates or interim releases to Windows. And there's the sense that maybe some companies that bought Software Assurance feel empty handed for products that aren't going to make it in the calendar year.

Ballmer: We have worked, we will work, very hard to make sure that anybody's that ever signed a contract with us to buy software feels like they've gotten good value for money paid. I make that as statement of religion. I'm not going to go down any specific path and announce something new, but I will tell you I spend a lot of my waking hours thinking about the topic of, how do we make sure, not only for new things, but for things that people have already paid, there's a fair amount of value delivered. I think we'll do OK on that dimension.

From a road map perspective, I think all good road maps have to have an element of incremental advance, plus boldly going where no man has gone before. Because if everything is done incrementally it turns out you never get a big advance. It's nice, predictable, regular--a lot of things that people like--but there's some things that don't get done well on that schedule. We need a mix of incremental, that can be done on some more regular, predictable enhancement schedule, whether that's six months or a year or two years, and then we need to have bigger breakthroughs that may take more than a couple of years. Now, it turns out we're sort of in the middle of a cycle where we've got a lot more of the big advances coming--that's pretty exciting. But that has caused us to have a little bit less on the incremental, new-release perspective than perhaps in a perfect world we'd like. We're taking a look and thinking through some of those questions now, and seeing if there isn't a better way to skin the cat. Are we fundamentally rethinking our strategy? Probably not.

We are taking a look at the specific case of Software Assurance--not taking a look, we have made changes to Software Assurance to make sure there's more regular value that gets added independent of whether we're on an incremental release or big release. That's why you saw us introduce a set of incremental Software Assurance benefits last fall.

InformationWeek: What can you tell us about when to expect Longhorn?

Ballmer: We set ourselves a set of bold objectives for Longhorn. We went through the initial phase of trying to get feedback on that, being very bold, not crazy, but very bold. We got a lot of great feedback and we learned two things through that process: one, we have more in our initial Longhorn conceptualization than we probably even need to get the breakthrough result we'd want in the market. That's good news. And No. 2, we've probably got more work than we wanted to get Longhorn done on the kind of reasonable schedule we had in mind when we first started talking about Longhorn. So I think the key process which we've been going through is how do we make sure that we take advantage of this situation to cut back a little bit on the boldness--and still be viewed as a big, important, bold release--and get done on a more short-term time horizon. And I think we've made some pretty good progress on that internally; there's still some work that needs to get done. And the Longhorn release that comes out will probably still be a little bit later than I had hoped initially. Not as late as [it would be with] some of the all-singing, all-dancing stuff, but still the most important, significant release we've ever done.

InformationWeek: What about the so-called interim releases or updates to Windows that have been introduced in recent weeks as a possibility?

Ballmer: I'm not sure how much of that we've done and how much of that the press has done, frankly, on our behalf. There are things that I never thought of as a Windows release that I can read about in the press these days as a Windows release. I read one thing, I said, 'Geez, unless somebody at Microsoft said something really nutty, these journalist guys are managing now to make up releases.' They're code words for marketing campaigns; we're talking about marketing campaigns and somebody thinks it's a release. I've seen some pretty goofy stuff out there. I think it's well known we have some incremental releases coming of Media Center and Tablet PC, pre-Longhorn. We've got Longhorn. The real question has been on the server side: Do we go right to Longhorn or is there an intermediate milestone? Server releases take longer than client releases because we give them additional kinds of stress tests [that are] appropriate for that kind of server enterprise environment. I think the real question's been much more on the server side than on the client side. And I don't know where we are on the communications process exactly.

On the client side, you shouldn't expect anything outside of Media Center and Tablet. On the server, it's still a little bit up in the air. We may do something that we think would be a kind of incremental [release].

Linux
InformationWeek: Let's talk Linux and open source. What's your view on the new breed of commercial software companies that are based on open-source software.

Ballmer: There are two separate questions in there. One is, how will open-source software do? How are you competing? The second is, how will those companies do?

My answer to the first question doesn't vary no matter what you think the answer to the second question is. Open source is a fact of life; it's a kind of development that will be done. We think we have all the opportunities in the world, and will do a very good job of putting together a compelling, high-value, low-cost proposition versus the open-source guys we compete with. You ask me why, I'll give you a hundred reasons why somebody should select Windows versus Linux, why someone should pick Microsoft Office versus Open Office. We have a value proposition that's based not only on capability, but on lower total cost of ownership. Some of the things that we can do because we are a commercial software company that can change the product, innovate, stand behind the product will be seen [over time] as advantages even in some areas that people might wonder today whether they're not advantages for Linux. We'll have a very clear case on better manageability and lower total cost even though we have a cost of acquisition. We've got that today and it will get better. The conventional wisdom will be that the lowest total cost of ownership is our stuff. Right now there's some discussion about that. I don't think there will be any discussion about it over time.

Our ability to stand behind our products as a commercial company that does take commitments [seriously] will be very important on the security front. Right now, we've got all kinds of security problems; we're the No. 1 target, the biggest bull's-eye. Bill Gates was doing a college tour a couple of weeks ago. The thing that surprised him most was, he expected the professors to give him a bit of a hard time on security. He said it was quite the contrary. They said, 'You guys are doing some of the finest work you've ever done in the security area. And, frankly, you're going to know so much more than anybody else in the world about security because nobody else lives in the kind of environment that you live in with this many people trying to come after you.' In some sense, I think we're going to turn security over time--we have our work to do, but we're certainly putting the IQ, the energy in--to make that a real source of advantage.

If the question is, what's up with you and open-source products like Linux and OpenOffice? I say we're doing great work, I love our value proposition, I think it's stronger today than it was a year or so ago and it's going to continue to improve.

The second question could be, what do you think of these companies that grow up around open source? I think those are weird companies. I don't know whether they have a business or not. I won't say yes, and I won't say no. I just don't get it. They're not intellectual-property companies, so they don't have a business model that looks like a software company because there's no place for intellectual-property rights in the open-source community. If you make enhancements, you've got to give them back to the community. There's no proprietary, protected intellectual property. You already have kind of the Unix syndrome with Linux today: Linux is not Linux is not Linux. As that problem gets worse, our value proposition gets stronger because Windows is Windows is Windows. Part of what we went through in court was so that Windows could be Windows could be Windows. So, there's not a big value proposition for these companies to make it as standalone software companies, so maybe they're services companies. But they're weird services companies because they can't even commit that they're going to change the software. Red Hat doesn't control Linux. They don't control the road map, don't even control the fixes. Same thing with Novell and SuSE. Same thing with IBM. That is IBM's new model, they want to be some kind of packager of open-source software. Five years ago that wasn't the case. Today, you could say, where are they going to go with WebSphere? You can't love Linux and not love JBoss. Where do they draw the line? InformationWeek: What about companies such as JBoss Inc. and MySQL AB, which would claim that they control the code base?

Ballmer: Those companies are more interesting. But in a sense, they have some of the same problems. If they're open source, they're open source, and if they're not open source, they're off on a bunch of proprietary extensions. If this is just a question of saying, here's my free piece, here's my pay piece, they're going to wind up pretty quickly in the position of having to say, and my free piece isn't very good. It's a very tough set of economics. We don't have free software. We're not choosing between zero and something. These guys are choosing every day between zero and something. And it's not like they're huge companies and that's an easy set of decisions for them to make. So, how will they do as companies? I'm doubtful. I'm trying to be less doubtful about the technologies [than] I am about the companies. Because I think we're going to be competing against open source even if we're not competing against these so-called open-source-based companies. If you ask me to make my prediction, five years from now it may not be JBoss Inc. and MySQL AB and [others], but we're still going to be competing against the community that develops software.

Software Repositories
InformationWeek: Our publication recently reported on state governments, including Massachusetts, that are creating a repository of open-source software to share among local agencies, and some federal governments are developing open-source policies and opening labs. What is your reaction to those types of efforts?

Ballmer: I don't understand why any government wants to waste its taxpayers' money. I think it's unconscionable. That's different from saying, 'Somebody else's software, including open-source software, does a better job of meeting their needs than ours does.' Do what you think is right. But as a position of policy to say that government should be involved in the creation of products that compete in the software market? Come on, there's not even enough critical mass. The state of Massachusetts, a very important state, doesn't create a market phenomenon. Whatever market phenomenon exists, exists, and its chances of success aren't appreciably helped or hindered by the state of Massachusetts. I think the state of Massachusetts is best served by picking the products that deliver the best total value for the people of Massachusetts. If you take a look and say, how widespread is the phenomenon you describe, I think you overstate it. In this country so far, I think we have only one state on the record, and it's one very interesting state. They're on the record on something else I pay attention to, too. [Editor's note: Massachusetts is pressing for stricter restrictions on Microsoft than specified in its 2002 settlement with the Justice Department.]

If you go outside the United States, you do find some interesting phenomena. First of all, these aren't mostly projects supported by national governments. National governments are mostly following what I would call sane policies, that is, they're going to be good users of technology, wherever that happens to take them. You do get a few local/state-type governments that want to get out there, get some attention, some press. We've seen some in Spain, some in Germany. But the fact that we can name on our fingers [only] a handful of state governments around the world that have made statements around open-source software just proves how small a phenomenon it really is. Governments are conditioned to doing tender process, which focuses a lot on absolute price, and so one of the things we're having to work on with governments is to look at total cost as opposed to price. So, it's likely we'll see zero-cost-of-acquisition alternatives get more of a look-see, particularly in local governments, but at the end of the day, I think even those governments will end up making total-cost-of-ownership decisions.

In the case of China, China's always wondering, are we going to have our own standards or are we going to follow world standards? And it's one the few countries that's absolutely big enough to go through that kind of a process. We had some problems a couple of years ago getting off on the good foot with the Chinese government. I think we're well past those. We have a very constructive relationship. We just brought aboard a new president [Tim Chen], who used to run Motorola China and is very well connected in the Chinese government. And I think the Chinese are getting more and more comfortable with Windows as part of their infrastructure. They were worried at one point, will this have issues of national security? So we worked to make sure that the Chinese government had a [Windows] source-code license, so they can guarantee there weren't any back doors or inspections, appropriately, from people outside China. Last week, we made a major announcement with the [China Ministry of Information Industry], the major agency on this stuff in China, in terms of their support of .Net. That was a very big deal for us. So I think we're on a very good path with Windows in China.

In fact, our market share relative to Linux is better in China than almost any other country in the world. It might be our strongest country. You might say, hmm, that's odd. But countries that have high piracy, you could say in a country like that, our acquisition cost is the same as Linux. I don't mean to sound facetious, that's not where we want to be, but really, for most people in China, Windows and Linux cost the same amount of money. I was in a book store when I was in Beijing last time, and I see all these copies of Windows in a funny-looking package for $2.50. Then, way in the back, behind a lock and key, I actually found a Microsoft Windows for $99. It was our copyright--you couldn't see it, it was collecting dust, it was behind lock and key. The $2.50 version was the popular version.

Photo of Steve Ballmer by Rudy Archuleta/Redux Pictures

Return to the story: Up For A Fight