Ballmer learns from past Microsoft missteps

Steve Ballmer is remaking Microsoft in his own image. Aside from hiring on new blood at the software giant, he said some recent executive departures weren't voluntary.

7 min read
SAN FRANCISCO--Steve Ballmer is remaking Microsoft in his own image.

Microsoft's chief executive told CNET News.com in an interview here this week that while some of the recent executive turnover at the software maker was through retirement, some of the departures were not voluntary.

"Some have left, and I'm sad to see it. Some have retired. And for some, it was time to move on," Ballmer said. Since he took over as CEO from Bill Gates in January, Ballmer is surrounding himself with a clutch of trusted executives--some Microsoft veterans, and some new blood.

Ballmer is also beginning to learn from Microsoft's prior mistakes. In particular, he said the company would have done better moving from Windows NT to Windows 2000 in a series of smaller steps, instead of one big one.

In the case of Windows 2000, we made too many decisions of the things we wanted to change...we should have kept ourselves in better check. "Even though you're making the product better, sometimes it'd be better to ship less sooner and then come back with a later release," he said.

This gradualist approach is the one Microsoft has embraced as it moves toward its new .Net strategy, a new focus that contrasts with its current PC-centric business style. With .Net, software will be thread across a series of connected central servers, so people can access the programs through a PC, handheld or any other computing device.

The server versions of Windows 2000 exist at the deepest level of the new strategy. Microsoft this week debuted the top-end Datacenter version along with a collection of software packages for running email, databases and a variety of other server tasks.

At a launch ceremony in San Francisco this week, CNET News.com spoke with Ballmer about the rival Linux operating system and his disappointment over the company's WebTV Internet service.

CNET News.com: Can you give us an update on the state of the company? In light of all the recent leadership changes and executive turnover, how has the personality changed? Have you refashioned the company in your image?
Ballmer: Oh yes, in part. We've changed executives. Some have left and I'm sad to see it. Some have retired. And for some, it was time to move on. We have a different team, but we have an excellent team. If you look at guys like (former Silicon Graphics chief executive) Rick Belluzzo, (group vice president) Bob Muglia, (senior vice president) David Cole, (senior vice president) Brian Valentine?These are guys I've had a chance to work with close to 10 years. I have incredible respect for these guys and they're making it happen everyday. I'm excited about the management we have in place and we're still adding to it.

What has Microsoft learned from the development of Windows 2000 and the long delays in releasing the business operating system?
We learned, in another way, lessons we've learned before. The toughest thing in software is knowing how to change it enough, and not too much. Every decision you make is for the good, but if you keep changing what it

Click here to Play
Datacenter debut
is you're trying to build, you're just pushing the development time of a given release too far out. Even though you're making the product better. Sometimes it's better to ship less sooner and then come back with a later release.

In the case of Windows 2000, we made too many decisions of the things we wanted to change?'Let's just add this. Let's just add that.' And we should have kept ourselves in better check. We should have rolled out a version without the directory and without some other capabilities. We did a good job, but we're never complete, and you shouldn't always try to be complete.

Next: Sales, and other software concerns 

Sales, and other software concerns

Analysts are expecting slow Windows 2000 sales through the first quarter of next year. What's going on? Are people just not rushing to upgrade? And what's the revenue impact?
Windows 2000 is doing fine and could be doing better. We're meeting our Ten years from now, more of our revenue will come from (software) subscriptions. business goals at the server level. The client level could be stronger, but it's actually meeting our plans. It's not meeting our dreams, but it's meeting our plans.

Windows 2000 is a business operating system, not a consumer operating system today. It's not until "Whistler" next year that it becomes a consumer operating system. Because it's not a consumer operating system, you don't get the mass end-user upgrades. You don't get end users walking to their bosses in the office, saying, "I must have this in my work." So, there's some of that end-user enthusiasm bootstrap that we never had with Windows 2000. But we do see a lot of evaluations and expect a lot of deployments in corporations this year.

The big issue right now is not Windows 2000. It's a question of what's happening to the PC market. The PC market is very large -- about 130 to 140 million units. It's growing, but not at the same rate as it used to grow--at 25 percent plus. It's a little saturated, but there's plenty of growth in the old guy yet.

Do you expect Microsoft's revenue will be bigger with the high-end Datacenter version or with the PocketPC/Windows CE operating systems for handheld devices?
For the next five years, certainly the Datacenter stuff. In 10 years, it's possible they will be equivalent.

You've said that the subscription-based model for software will someday make up a huge chunk of Microsoft's revenue. When is that point coming when subscriptions over the Web overtake traditional software sales?
Over time is the answer. Ten years from now, more of our revenue will come from subscriptions. When is the crossover point? It's not in the next three The big issue right now is not Windows 2000. It's a question of what's happening to the PC market. years, probably not in the next four years. So that gives you a range of four to 10 years. It's inevitable. Software that takes care of itself, backs itself up on your behalf, stores it for you on the Internet--how could you not prefer that to software that doesn't do those things? That's why I think it's inevitable. That's why we have to build the platforms to support it. We will have some applications that we will never do in package form. If you look at our bCentral Web site (for small businesses), you'll see services that we'll never deliver in package form.

There will probably be a day three years from now where we're saying, 'There's not enough bandwidth in the world' (for the subscription-based model). And customers don't want this stuff. We'll have to be patient.

What Linux efforts are underway at Microsoft? There are persistent rumors that you are creating a Microsoft Office for Linux.
We certainly have no plans. Linux is not catching on, on the desktop. There are no customers. I may be from Mars, but if there's no demand, we're not going to do the work to take Office to Linux. It's not even an interesting question until there's demand. Linux on the server is a different story. We might still dramatically outsell Linux on the server. You don't see much Linux in (business) customers. You see some Linux in Web sites and application service providers, but it's less than the press hype. You still see a lot of Linux in Internet service providers, in academia and in the embedded market. We have geared up on our embedded efforts and formed an embedded division. It's an important thrust for us and we've started to win important embedded design wins.

WebTV seems to be stuck at one million subscribers. Is it PC saturation? Or the lack of interesting applications? Why is it seemingly going nowhere?
Yes. I actually think that the current WebTV service is interesting but not overwhelming. What we do today is let you get Internet on TV, as opposed to enhancing the TV experience. We recently introduced a service called ultimate TV, which is much more about enhancing the TV experience. I do expect we'll see growth of subscribers for UltimateTV. Frankly, the No. 1 reason people churn out of WebTV is because they have enough money to buy a PC. UltimateTV will be different. You get the Internet on TV, but it's more about enhancing the way you watch TV, recording shows, pausing. I'm bullish on that.

You've compared .Net and Web services to the notion of VB controls, OCXes and ActiveX controls. They were the building blocks for Windows programmers in the '80s and '90s. Web services, which can be reused by Web-based subscribers, will be the reusable component in the future, you've said. But unlike those earlier components, Web services won't necessarily be tied to Windows. How does Microsoft simultaneously spread the Web services notion and keep a big piece of that business for itself? Where's the proprietary advantage?
Web services speak XML (a Web standard for exchanging data). So they can run on a variety of different places. Our job is to make sure we provide the best tools to develop and deploy Web services. And if we do that, we'll make business. If we don't do that, we're not going to make business.