Our Sit-Down With BallmerOur Sit-Down With Ballmer

IW: You're pushing the business value of Windows XP, but our latest research shows more interest in Windows 2000 desktop upgrades than Windows XP upgrades. How do those two platforms compare in terms of current shipments with business customers?

Ballmer: Windows XP would be bigger, I think, than Windows 2000. I think that's right, I haven't parsed the numbers that way. But it's not like 90/10; they'd be close. What happens is you get customers who standardize on something like Windows 2000, and then once they standardize, they're going to want to continue to take Windows 2000 even if Windows XP is a bit better--and Windows XP is a bit better. The big focal point we have these days is on customers who are still on Win95 and Win98 and the oldest versions of Windows NT workstation. For those customers, it represents such a compelling value proposition to go to Windows XP and Office XP, that we just need to get that word out. We're a year after the launch of Windows XP. We've issued our first service pack. Whatever concerns people might have had about a new release should be shaken through the system. There's been time to work that out. We've had a number of major customers now make the switch from Windows 95 or Windows 98 to Windows XP. What we found is, on average, people seeing something like $350 to $400 savings per desktop over a three-year period. We set aside 1 million bucks and we said, look, the first 40 customers who have 5,000 or so 95/98 desktops, we'll run it and prove it with those guys, that they're going to see that kind of cost savings from the migration. And that's net of the expense to do the migration, frankly. So we feel pretty good about that overall value proposition around the Windows XP, Office XP desktop. We don't have a big push relative to the Windows 2000 customers. XP is better, but the improvements aren't as dramatic as they are for customers coming from Windows 95 and 98.

IW: And what's driving those savings?

Ballmer: Reduced support, reduced support incidents, reduced help-desk management and maintenance costs, fewer crashes, better mobile support. Mobile are probably 25% to 30%, at least, of computers that go into larger enterprises--and the configuration hassles, the standby, the resume time, all that stuff is so improved.

IW: The counterpoint to this is the discussion that's still going on around your newest licensing policies. Some users feel their software costs are rising as a result. You hinted in public remarks Oct. 9 that you'll be taking another look at your licensing policies. What did you mean by that?

Ballmer: I didn't hint that, at least I didn't intend to. What I said was, we've made our changes, we're done, we're through that process, there are some things I'd do differently. The second thing I said was, I didn't think we would make any subsequent changes--we've been through that. The third thing I said was, we started this set of licensing changes to improve simplicity of licensing from a customer perspective because we had all these different licensing options, and that was an important simplification. Although there were collateral issues that we created, I don't want to create those again. I did say we have a bunch of people who still find our end-user license agreements overly complex. I said we would like to simplify those, too. It's not a price issue or anything else. We would like to simplify those also, but we have to make sure we don't make the mistake this time of starting out to simplify something and causing collateral issues where people are concerned about cost. The end-user license agreement is a document that says how you can use the product. Those are long and fairly complicated today. Many people will take a look at your end-user license agreement and say, "Am I allowed under the license to do this or not, and if I am allowed, what do I have to pay to do this?" And many people have a hard time because this is fairly complicated, not because we want to be hard to understand, it's just some of the ways the software gets used is pretty complicated. We would like to simplify that just like we wanted to simplify our volume license terms, which is the thing we simplified in the last round. If we can't simplify them in a way that doesn't create collateral issues, where people are worried about the price and cost of implementing our stuff, we're going to leave them complicated. If we can simplify them, and do that in a way that people can feel pretty good that we haven't raised their costs, we will simplify.

IW: When we've reported on some of Microsoft's licensing changes, it's easy for us to find customers who feel that their software costs are going up. It's harder to find those who feel that their costs are going down.

Ballmer: Anybody that was an Enterprise Agreement customer of ours who's renewing by and large sees their costs go down. We have plenty of Enterprise Agreements like that.

IW: Their costs decrease by how much?

Ballmer: Percentage points. It's not like a factor of 50%. Prices come down by and large for somebody who's an Enterprise Agreement customer, and they can go up for some other customers.

IW: What leads to the price drop for those that get it?

Ballmer: We cut the price. (Breaking into laughter.) It's not very complicated. We cut the price.

IW: But we did talk to some larger customers who at least thought their prices would go up. Is there some type of confusion about that?

Ballmer: Costs went down. Now, some people have elected, since an Enterprise Agreement was signed three years ago, some people are saying we've got more users that we used to, more desktops. Or people are saying, we'd like to take more of your servers. So their total bill to us might be down or might be up, but the cost per desk on that Enterprise Agreement for somebody renewing is definitely less than it was on those initial agreements three years ago. An Enterprise Agreement is something very concrete. Before, we had something like Microsoft's new Software Assurance, and the Enterprise Agreement essentially had Software Assurance built into it.

IW: Help us connect the dots on some of this week's product announcements: an upgrade to Exchange server code-named Titanium, an XML server code-named Jupiter, and an XML desktop tool code-named XDocs. And you've got Windows .Net Server coming up. How are these things related?

Ballmer: Loosely. Windows .Net server is a new server operating system. Jupiter is the next generation of BizTalk and some other technologies, but particularly BizTalk. BizTalk is an XML workflow engine and it's now programmable via .Net. It shares that in common with the Windows .Net server. Titanium is a next release of Exchange which I think people will get a huge benefit from. It's being released on top of the Windows .Net server. So, you could try to look cosmically for a deep connection--they share the same technology base and some of them have dependencies on others, but they're not 100% linked up.

IW: You mentioned in your public remarks today Oct. 9 that you want to help customers take advantage of the XML revolution.

Ballmer: Everything we do shares in the fact that we'll try to help people exploit the new lingua franca of the Internet, XML. XDocs does that. BizTalk and Jupiter do that. Certainly Windows .Net Server does that. Titanium has less in it, frankly, that is XML-focused in this release; it will have more in the next release. Mostly what Titanium focuses on is responding to the many desires customers have expressed for improvement in the Exchange 2000 server.

IW: What kinds of business processes do you see enabled by this XML revolution? And what role will Microsoft play in enabling them?

Ballmer: We call XML a lingua franca for the Internet. It's a common way for programs to talk to programs, people to people. It's a common representation type for information among computer systems, not just on the Internet. When people say, what's the future of enterprise applications integration, the answer is, everybody's going to move their architecture to XML. When people say, "What's the future of E-commerce?" the answer is that everybody's going to architect around XML. When people say, "What's the future of desktop application integration?" the answer there, too, is XML. My wife is a big XML lover and she doesn't even know it. She just keeps telling me she wants her MSN calendar and my Outlook Exchange calendar, she wants to be able to drag and drop one on top of the other. That's the enterprise application integration problem for the Ballmer household, and the answer to that problem is those things will be architected around XML. So it becomes the basis for integration and connection. .Net is a set of technologies we're infusing into Windows and elsewhere that help with that. And we're trying to improve in general the degree to which our development tools, our operating systems, and our desktop applications let people work with and manage XML information. XDocs is part of that. The .Net support in Windows .Net Server is part of that. What we've got coming in future releases of databases and the Windows operating system is all part of that. Our new version of Office, in addition to XDocs, will have enhanced XML support. We're targeting these scenarios where you pass and manipulate information in a much more reusable form across systems by representation in XML format.

IW: The other part of the question has to do with business process. It's a phrase we see popping up in announcements of products like Jupiter and XDocs.

Ballmer: Yes, those two would very directly.

IW: What is Microsoft's role in enabling business processes?

Ballmer: We see ourselves as a business-process enabler at probably three different levels. For a long time, people have used our tools to write business-process applications. They've used Visual Studio, they've used Office sometimes as front ends. That's very much in the blood, and we've seen that in enterprise customers and smaller customers. That's a pretty low-level set of tools to use and you've got to do a lot of work yourself. The second level up is a set of higher-level tools that help people literally orchestrate, plan, and develop business-process workflows. That doesn't mean you don't need to plug in code, but if you want to say, "I want to get this invoice from point A to point B," how do you describe the process and how do you describe what computation should happen against that invoice at various phases? XDocs helps with that. BizTalk Server helps with that today, and Jupiter helps even more with that. Enterprise application integration is kind of at that layer, because enterprise application integration isn't really about the computation, it's really about moving the workflow from compute station to compute station. That's kind of what BizTalk server does today. XML again is the engine. The third layer is business-processing applications themselves, which is what, at least for smaller and medium-sized businesses, we deliver through our Great Plains and Navision products. So there's three different layers at which we play today and we'll increasingly play as we go forward.

IW: You mentioned that your Great Plains and Navision applications are suited for small and medium-sized businesses. Going forward, will you extend those apps to larger companies?

Ballmer: There will be some people who call themselves enterprises who will be very happy with Great Plains and Navision, but our design point is for small up, as opposed to large down. If you take a look at the design point for SAP or Siebel, their target is bigger enterprises and they'll go down as small as they can. We're targeting smaller enterprises, we'll go as big as we can. There's a little itty-bitty overlap someplace in the middle where customers will have a choice. But for 90% of what they do, and 90% of what we do, there's not going to be any real overlap. They don't want the 10-person local manufacturing company and we're not going to pretend you can run GM on Great Plains.

IW: Are the Navision and Great Plains products going through changes to hook them into the IT architecture you're describing?

Ballmer: Yes. Navision and Great Plains are being re-platformed on top of .Net technologies and consequently, XML. This is a multiyear re-platforming. And when we do the re-platforming, we're also going to converge the two product lines. We don't need two sets of application logic. So we're going to take the work of the applications capabilities and kind of spread it out across the new integrated Microsoft Business Solutions team. That's a multiyear project we're moving vigorously on today.

IW: When we visited the Microsoft campus earlier this year, there was a strong sense among some of the people we talked to that Microsoft needed to reinvigorate businesses and knowledge workers around its desktop applications. How much more headroom is there for increased productivity in terms of what you're able to deliver to them?

Ballmer: Huge. Do you come to this meeting any differently than you would have come to this meeting 10 years ago? You're a knowledge worker. Have we helped you? No, we haven't helped you. Just take a scenario like this. And if you're honest, you'd say, if we don't provide tools that assist you in doing a better job in a meeting like this--this isn't a condemnation of you, this is just a commentary on the fact that the technology isn't helping you. You need the right hardware form factor with the right software, the right built-in support for audio, better tools for note-taking, better ways of organizing old archived information so you can rapidly get at it, better way of collaboration because I'll bet there's at least one thing I've said in here that you'd like to fire off an instant message. Where is it? Where is it? Maybe you'd like your notes indexed against the tape so that when you say, what was Ballmer saying when I made that note, you can do that. You can have a complete capture of this meeting. Just a few things that we're not doing for information workers. I could say the same thing about document sharing and information sharing. There's a slew of things we're not doing. What about business intelligence and information analysis? There's a ton of things that we're not doing.

IW: Some of the potential is pretty obvious, but is there enough value to get businesses to spend a lot of money on these things?

Ballmer: If businesses think there's a return and they're interesting, and the users think they're interesting, they'll absolutely spend some money. The fact of the matter is, if you take someone whom you pay $40,000 a year or more, and you say, I've got to spend a few thousand dollars a year to make that person 30% more productive, it's not a hard payback equation. Relative to the number of hours you work, if you can get a little bit more productive because information technology is helping you every hour, that's a pretty good value. This isn't just a value proposition from Microsoft. This is what the whole IT industry has as its opportunity. People say the big wins in IT are the backroom. Well, some of them are, but when I meet with CEOs, the things that really get them charged up isn't me telling them, hey, here's an SAP implementation, it's hey, here's how you get access to information that lets you make smarter decisions. Here's how you see profitability by account, here's how you connect your suppliers and actually have them have real-time visibility on your data and vice versa. Our industry has to deliver that kind of compelling value, and Office just happens to be kind of out there at the front end where the users use the data, which is a pretty important place to be. But that's the value chain that the IT community is all working on. It's not just about cost reducing what we do today that's important, but it's also about helping empower people in new ways.

IW: Jeff Raikes has talked to us about delivering new business-intelligence capabilities to Microsoft customers. What can you tell us about that?

Ballmer: Some in Office 11, some over time. We think can do a lot more to make numerical data dissectible and understandable by people. We think we've got some pretty interesting ideas in that area.

IW: Last question: Who's going to use XDocs?

Ballmer: The average worker will use it, empowered by scripts, workflows that are created by super users and IT people. Mostly it will be the end-user interface on a set of business processes that maybe a super user but typically an IT person would have created for them.

IW: Do you have something in that space now?

Ballmer: No. For some of that kind of stuff, people will use Excel or Access. It's not perfectly tailored for it, but it's some combination of processing engine and forms engine, and people will use Excel and Access for some of those kinds of things even though they're not as XML- tuned today, they will be, as XDocs. XDocs is really targeted at the business-process flow and the end-user interaction. It doesn't have other purposes the way Excel or Access do.

IW: So it won't be a personal-productivity tool launched by every person that sits down at the PC in the morning? Ballmer: Oh, I suspect it well might be, but it's one that gets launched because you launch the application that uses it, as opposed to saying I'm going to launch XDocs to create my own work raw. There are two models in Office applications in general: one where people are manipulating the work of others, and one where they're creating their own work. Even with PowerPoint, many people start with a presentation that was done by somebody else and they manipulate it, or an Excel spreadsheet that was created by somebody else and they manipulate it. I think that will be the most common form of interaction with XDocs. Some people start with Excel and go build a worksheet, some people start with Word and go build or type a document. I think that will be less of a common starting point for XDocs.

Return To Road Map Redefined