A Sit-Down With Steve BallmerA Sit-Down With Steve Ballmer

Following Microsoft's April 24 launch of Windows Server 2003 at San Francisco's Bill Graham Civic Auditorium, Microsoft CEO Steve Ballmer talked with information senior editor-at-large John Foley and senior writer Aaron Ricadela. Here's an excerpt from their discussion.

information: You said in your keynote presentation that you weren't sure what to expect in the way of Windows Server 2003 adoption, given the economic climate. But what's your best guess?

Ballmer: We don't do much in the way of forecasting. The thing I wondered about--I know we have a great product, I know it has a lot in it, and I thought, based on the feedback we had from customers, there would be a lot of enthusiasm. The question would be, How much? What would really be the case? There's this general sense when you talk to customers and to journalists and analysts, 'Oh gosh, everybody's sort of in the doldrums.' To have 280,000 people signed up to come to 170 [Windows Server 2003 launch] events--we're not talking about Windows XP here, we're not talking about a client-side product, we're talking about a product that is relevant to IT people. It's a stunning, stunning, stunning show of interest. The devil's in the details, we'll see how things sell, but I am heartened by the level of interest in this product. And I think it is because the product is the right product, as it happens, at the right time. If you're in the middle of a boom, go-go, everything's great economy, it's still the same product, but it might have generated less enthusiasm because this whole notion of not only more, but more with less, is so much more important at this time.

IW: Could you give us a point of comparison on how you expect Windows Server 2003 adoption to compare to Windows 2000?

Ballmer: There's two interesting dimensions. Of the new hardware servers that go out in the next 12 months--that would be 5 million or so servers--what percentage of those will go with Windows Server 2003? What percentage of them will go with Windows 2000? And what percentage will go with non-Microsoft? Those are interesting questions. I think that we should do significantly better in terms of new server run rate with Windows Server 2003 than we did in the first year of Windows Server 2000. Part of that is because we don't have as much of a disruptive technology--Active Directory, which came with Windows Server 2000, was more disruptive. You had to do more thinking, more planning. This has a lot of great stuff, but more of it has the characteristics of being responsive or additive, as opposed to disruptive. Disruptive's not bad; it just takes longer to adopt. The .Net Framework--that's additive. Windows Sharepoint Services are additive. The improvements we've made in security, reliability, manageability, those are all customer responsive.

Another interesting question is what percentage of Windows 2000 servers, and Windows NT servers, particularly NT 4--those are getting longer in the tooth--at what rate might those upgrade? Will they upgrade the software? Or will they be both hardware and software upgrades. That's one that's harder to predict, not because it's not a great product, but because IT professionals do occasionally adopt an attitude that if it ain't broke, don't fix it. And although the NT servers are doing a fine job and while this might be better, in this climate some of those might just get left alone. So, I think we've got the right product for that upgrade. I'm enthusiastic about that upgrade, but we've got to see what the market really wants.

IW: If you were to identify one thing that you think will drive adoption of the new operating system, what would that be?

Ballmer: Better basics. Improvements in reliability, security, performance, scale, manageability. Those are non-scenario specific. The foundation is better. That would be number one. Number two, I would probably point to the fact that it's a better file server, because of the collaboration infrastructure and Sharepoint Services. Those are the two things I would point to that will drive the most volume. Now that may sound unusual because I was out there [on stage] hugging the HP Superdome and high-fiving [Intel president] Paul Otellini, but from a unit-volume perspective, if we knock the ball out of the park with that thing, we're talking about thousands of units. We're not talking about hundreds of thousands or millions of units. That market is important from a unit-volume perspective, it's just smaller. So, when you talk about mass adoption, I've really got to point to the basics, which impact everybody, and the single biggest market segment, which is that file server market.

IW: Is there going to be an obsolescence factor that's at play, not just that Windows NT servers are getting older, but that the support cycle is going to run out on those systems?

Ballmer: I don't think that will be the number one driver. We've extended the support time frame for Windows NT, and I don't think that's going to be the key driver.

IW: The flip side of our question is, what is the biggest factor that will hold adoption back?

Ballmer: It's a great release, we have more customer experience, longer beta cycles, more production deployments, more joint-development customers than we've ever had before, but there's still this natural, human thing that happens when people say, I think I'll wait for the first service pack. There will be some of that. That's probably the thing I would point to as most significant.

IW: Rather than the IT resource issue?

Ballmer: Well, people like to talk about that in a funny way. The server market, by all forecasts, including our own, will grow the next 12 months over the last 12 months. So, budgets may be tighter, but that doesn't mean fewer servers are going to get sold. As long as more servers are getting sold, there's more opportunity to sell Windows Server on top of it.

IW: In a survey done by information Research in January, approximately 29% said they would adopt Windows Server 2003 in the next 12 months. Is that in line with your expectations?

Ballmer: There's two different things. Is it in line with our unit-volume expectations? And is it in line with what I would expect people to answer in the course of a survey conducted before the product ships? It is in line--actually, frankly, kind of high for a survey done in a tough economic climate for a product before it shipped. I think [laughing] it's a pretty good number. In terms of unit volume, if you asked me to forecast the five million plus servers, at least the ones that ship with Windows, I think Windows Server 2003 is likely to, by the end of the period, on a run-rate basis, I'll bet we'll be getting on towards half. You might say, half, why only half? There's a lot of inertia--people haven't tested their applications, they're still waiting for a service pack, they're using Windows 2000, whatever the case may be. I think it could be half. I don't know if that's right. I'm not staking my life on it. It's not a financial question. It's not going to affect our earnings; the price of the new version is the same as the price of the old version. But I think particularly in smaller and medium-sized accounts it will happen quite quickly. In larger accounts, there will be a little more time people take to test and assess.

IW: You called Active Directory a disruptive technology. Is it still presenting a challenge?

Ballmer: At this stage, people have done a lot of planning, and they've got skills built up around Active Directory. And we have really taken the tools to the next level in Windows Server 2003 in terms of setting up and managing Active Directory installations. So you get the combination of skills build-up in the customer and the partner base and the fact that we've made things easier and given them better tools--I think we're over that hump.

IW: Licensing and the cost of Microsoft's software was an issue in the second half of last year. You've made changes to the client-access licenses and held prices flat in terms of the operating system. Do you feel you've adequately addressed your customers' concerns about the licensing issues? Are any changes in store?

Ballmer: We tried to do a lot to address customer concerns and we're always trying to be smart about how to do better. The biggest issue our customers had was the fact that Licensing 6 was just so different than that which preceded it. Yes, there were some specific issues that we had to address, but I can tell you why I think that Licensing 6 is a lot simpler than Licensing 5. But if you knew Licensing 5, then Licensing 6 just looked different to you, and that caused it to be harder and confusing and everything else. Once we got through that, and there were specific issues we needed to address, we're in much stronger shape. It's important for us not to make major changes going forward; that is part of what we learned from the transition to Licensing 6. People don't like to see these major transitions. Our move to Licensing 6 was not driven by a desire to make a lot of money, it was actually a desire to try to simplify some things for customers. We learned the pain of that was not worth the gain, from the customer's standpoint. So you can expect a lot of consistency. That doesn't mean we're not going to continue to tweak things and be responsive to the kind of input we're getting from customers, as you've seen in the client-access licenses, the per-CPU changes we've made, the pricing.

IW: During your recent earnings call, CFO John Connors said Microsoft will be looking in fiscal 2004 at the total value proposition for Software Assurance. Can you explain what he meant by that?

Ballmer: We will explain that. One of the things people might want to see us enhance is the level of predictability that comes when you sign up for Software Assurance. We think we have a pretty good proposition today, and we think we have some pretty clever ideas to enhance it, to deliver customers more value and deliver it predictably as part of a Software Assurance agreement.

IW: On the subject of predictability, let's talk about the way that Microsoft is delivering technology as a part of this platform. There's a list of add-on software that will be rolled out over the next three to six months, but it's not clear which might cost extra.

Ballmer: There are different ways we will extend [Windows Server 2003]. One is with technologies that are essentially part of the product, but which may not have been completed at the time of shipment, and therefore will be available as downloads to any customer who's got Windows Server 2003. I'll call that part of product. Then there's a set of technologies which we will provide to customers under Software Assurance. You could say they're not part of the product, but they are part of the Assurance proposition. The third category is separate products that have a separate fee that work with Windows Server 2003. For those products, there's still a question for some of them, particularly for the client-access license, whether it's all separate or whether some of those technologies might be included for a customer who's got an Enterprise Agreement even if they are separate products with separate fees for customers who do not have an Enterprise Agreement. Around the base [operating system], you'll see us bring technologies out in all of these buckets. SQL Server, for example, is clearly a separate product. Windows Sharepoint Services we clearly view as part of the Windows Server product; they will ship a little later, but they will be provided to all of our Windows Server 2003 customers. We haven't even decided on everything, so if you ask me about a given product, I may or may not know, we may or may not have decided, and I may or may not tell you.

IW: Is this consistent with how you've been doing things until now, or somehow different?

Ballmer: Consistent. Servers are a little different than clients. In clients, people don't really want to stream a lot of new technology down all the time because software developers really need a fixed point to build on. The client you tend to upgrade all at once. In the server space, the IT professional has a greater demand for ongoing, sustaining improvement that is additional capability that adds value to the basic offer they bought into. So there's more of this kind of activity level required on the server than on the client, and we think that with Windows Server 2003 we have an architecture and a platform that will let us deliver more of this add-on value out of cycle, or out of band, as we say, with major releases. Let's assume the next release of Windows Server is at least two years from now--and it is--is there any innovation that's going to be provided to those customers between now and two years from now? The answer is yes. In the server case, people don't want to reinstall the server operating system, they want those things to be part of the system, but install on top of the current base that they have. There's a greater need for this stuff, and you'll see more of it from us in some areas than you have in the past.

IW: Is the out-of-band approach new?

Ballmer: We have done some of it. We are accelerating the pace. There's a set of ongoing innovations that we've delivered; we will step up the pace of innovation between major releases versus where we've been in the past.

IW: Let's talk a bit about the value proposition. You mentioned earlier that Windows Server 2003 can lead to a 30% reduction in certain IT costs. But our research shows that only 28% of your customers expect Windows Server 2003 to result in lower total cost of ownership. Do you feel customers have a clear equation for the TCO benefits?

Ballmer: With the testing and analysis we've done, and customers we've worked with, that's the kind of improvement we're seeing. I'll bet some customers will see much better numbers, and some will not see a number that's quite that high. I'm not going to say it's one number that fits all. But I stand by the fundamental assertion that we will help our customers do more with less, in part by taking out cost in the IT department. I feel very strongly about that being a critical part of the value proposition for this product.