No-Holds-Barred Ballmer

On the heels of Microsoft's dramatic "Tailwind" sales reorganization and updates to its core partner program, CEO Steve Ballmer met with CMP Channel Group President Robert Faletra and a team of editors to share what's on his mind. An upbeat Ballmer addressed Microsoft's vertical segmentation, competition with Google and other market segments that the Redmond, Wash.-based software company plans to study carefully during its upcoming fiscal year, which starts July 1.

CRN: There's been a long, industrywide discussion about managed services. You have the OneCare security service you've announced, which is supposed to be at the consumer level. But small businesses really seem to be served by the offering, too. What's the strategy long term? Will Microsoft be in managed services?

BALLMER: I'll tell you what I do know. Let me draw a picture: The new view of the marketplace. Forget small, medium, large. Too confusing. Let me make it even simpler. You have four kinds of businesses: very large businesses, and--I will define these in a different kind of way--midsize businesses, small and tiny. These aren't official Microsoft words, just a concept. Let's say tiny is a company that has no servers. It has a different IT environment. And think of a small business, just for kicks, as one with just one server. And a midsize company has a few servers and a few IT guys--maybe up to 10, and it could be 11. They might have someone who thinks of himself as IT. And large is everything else.

What you might ask is what each of these constituencies might want. The tiny business is a lot like a consumer. A family has four or five members. Maybe they all want one [Web] domain, and they need a common security contact. A family is a lot like a tiny business. Whatever we say is good for consumers is probably just as good for a tiny business. That's just my opinion. Now people might decide to morph from tiny to small or small to tiny. You might have people who don't have servers [decide to] put in servers for a variety of reasons. It doesn't mean the business size has changed, but they have gone from one IT characteristic to another. And you might find people who have a server today decide they don't want to have a server.

A midsize company clearly can take a bit more complexity, but not a lot of complexity. Say they have five, six, eight IT guys. You can't afford to have material experts and 28 different technical topics. You need something that's relatively simple. And a large business has its own set of characteristics. It may have a lot of capabilities, but whatever it does for itself or has done for it, it has to fit in its own concept of security and management and directory policies.

Why do I say this? I think that the strategy for us and our partners together should be to think through how to take costs out of some kind of designed offering at each level. Some strategies may work at some levels and won't work at other levels. Certainly, with the strategy you use for someone who's tiny, it's easy to see that it probably won't work for someone who's not tiny. Someone who has a single server has an interesting point; you have a little different strategy. Whatever you do is a lot more complicated for a midsize customer, but that midsize customer is not large enough in the number of PCs to get a lot of customized attention for whatever should exist for services. And a large business will have its own set of complexity. So we are sitting here saying, across the landscape, what do you want to do? In a tiny business case, [would they use the] OneCare? Sure, we will sell OneCare to a tiny business. There should be no confusion about that.

Whether there's a heavy channel play in a service perspective, I don't know. Probably not. There probably is a resell play because our channel partners do a lot of different things. Sometimes resell is the play, sometimes service and sometimes both. In small businesses, so far we haven't come up with any better strategy than to design serviceability into our Small Business Server to make it easy for our partners to participate. Now over time, maybe there are some services that today people have to run themselves that you could run for them someplace else. There are third parties that do this very well. There are people who say, 'Send me your IP traffic. And I'll filter it for viruses, I'll filter it for spam.' You could call that a managed service. It's not managing everything. But it still comes out of the data center, and there are still partners involved in the account.

That's quite different from OneCare. OneCare assumes you have no personality of your own except what you get out of our data center. [My broader philosophy] says you do have a personality of your own but that maybe there are some services you can farm out. For midsize companies, I don't know, exactly. We are trying to think through what we need to do with our whole product line to have them be better optimized. ... CRN: Let's talk about the midsize server concept. I was told there would be a midsize server?

BALLMER: If that's what they told you, that's what they told you. Frankly, I don't really know. Do I think there's an opportunity to innovate, engineer a product or product line that has greater capabilities than Small Business Server that lends itself better to the needs of an IT shop with five to 10 guys, plus support from partners? You really have to think about that whole ecosystem. If you only have five guys in an IT shop, you can't even hire that many partners because there's not enough of you to run around there to manage this work that's coming out of partners. So this is an innovation opportunity in the product line, and maybe in the service. There are some services that probably make sense in small to midsize companies. Those five- to 10-guy IT shops could want things like spam filtering and e-mail filtering. [With] really large companies, all we can say today is while there are some people who make good money at outsourced desktop managed services, there's been no breakthrough. It's brute force, wring some costs out and move on. That's the basic approach. It's not an engineered solution to take costs out.

We said to ourselves, 'Look, we have to engineer a solution.' Whether we deliver it, who knows? Whether our partners deliver it, who knows? Whether big partners deliver it, who knows? Don't know. No decision, no concept. But we'd better do the engineering, which means we can do the learning. Take a few customers, find out what they need. How do we engineer out the costs? We've signed a couple customers so far that we have deals with. Energizer is an example. The key thing is we ask ourselves what's the right way to take cost out through service-based offerings? Whether we offer them or our partners offer them, or some combination, it partly depends on what you engineer. It's not one of those things where you can just go into the labs. You actually have to work with the customers.

This, we're learning. Here we are working on the servers more than the services. Over time, if you were to say there's more software that's going to be provisioned out of central data centers than today, what do you call Windows Update? That is a large managed service. So are we going to invest in some of these things and try to innovate? The answer is yes.

CRN: And yes in the managed antivirus space?

BALLMER: Where you have a few IT guys, will we have a few products in AV? We said we were going to have an AV product, but we haven't said when. We haven't said whether it will be a service. ...

CRN: I don't want to spend too much time on Google, but a few questions. You say they're only doing search, but they're doing e-mail. They're looking for a head of channels, which would lead you to believe they're doing something they're not doing now.

BALLMER: Well, they have a search appliance.

CRN: They have a lot of money. The stock is at a some ridiculous multiple. They're probably just an acquisition away from getting into some other business.

BALLMER: Sure, of course they are. That doesn't mean they'll be successful. But they're an acquisition away. If money's the criteria, we're an acquisition away from anything. SAP? I heard the rumor. Oh, that was last year's rumor. Sorry. They have the money they have. I don't mean to be goofy. They've got a lot less money than we do--cash, market value, whatever it is they need to spend, they've got less.

CRN: Why don't you buy Google?

BALLMER: Some people would argue it's well-valued, some people undervalued and some people overvalued. There's not something there I want to buy. I'm not saying it's not a good company. It's a very expensive company. Do you actually believe there's not enough innovation in search that somebody's not going to be able to challenge them in search? It's the easiest application in the world to switch.

CRN: We're hearing of a sales reorg at Microsoft called Tailwind that will 'programize' a lot of the vertical stuff you've talked about for some time. Can you talk about that? How big of a change is that?

BALLMER: At the highest level, it's three things. No. 1, we want to improve our verticality. Saying [we want to be] more vertical is a funny word. Some things will be more, some actually less. This is U.S.-only. We think we'll be more effectively vertical than we've been in the United States. Goal No. 2, we're trying to get better integration of account management resources and what I would call specialists--specialists by vertical, specialists by technology. That's still a complicated thing in a structure that allows that to happen in more of a method, so that we can get deeper specialization in to work with our accounts but still have one sort of front end through the account manager--particularly in our biggest-of-the-big accounts, our enterprise accounts, as opposed to our corporate accounts. And No. 3, there will be some accounts that have had less account management resources that will get more, and some accounts that have had more resources that will get less. We'll rationalize that based on what the accounts have been doing with us in the last four or five years. Accounts go inactive for a year or two. But an account that has been less active for five years might get less time than an account that has been more active than we might have guessed, even though it has fewer PCs.

CRN: Do you characterize that as an evolution?

BALLMER: It's an evolution of solution selling in every sense.

CRN: Aren't Microsoft and IBM on the same track here?

BALLMER: There are things that are similar, but there are things that are very different. We're talking about selling technology specialties. We study what Cisco does, what IBM does. Why not? We are informed by that. IBM is very heavily verticalized. Given that the No. 1 thing they do is sell services, they ought to be heavily verticalized. I'm on the board of Accenture. Accenture is very heavily verticalized because that's the way services get sold. It's not the way technology gets sold. We need a nice mix. When you're selling management technology, it's not a lot different to sell management technologies to a bank than it is to a retailer or to a manufacturer. You have to figure out how to get the right blend, and we're tuning that blend, I would say, under Tailwind.

CRN: So we're hearing that a bunch of enterprise accounts will be pushed down to the small and midmarket solutions and partner group.

BALLMER: I don't know the number off the top of my head. They were tuning it [according to] the last review I had. Accounts that don't seem to need much or haven't been very active with us will get less, and accounts that are more active--and I'm talking about a five-year time frame--we will use the last five years to assess how active people want to be with us and how much investment we should make there.

CRN: This is a change, and that makes people nervous. Partners worry that their business will be disrupted while this stuff gets sorted out.

BALLMER: I understand that there's a sort of a tailwind of impact on our partners when we make changes, and the guys will try to do that in the way that's least disruptive as possible. It's the old expression, how do you change the engine while you drive the car? Carefully. We're not changing the engine, but we're definitely trying to do a little tune-up while we're driving. CRN: Microsoft Consulting Services (MCS) is another hot button with partners. We hear you're staffing up and that momentum is building for consulting services. Should partners worry?

BALLMER: Partners shouldn't worry. They shouldn't have worried last year, and they shouldn't have worried five years ago. They do. I'm not trying to deny the concern. Somehow we've been able to appear to change our strategy somewhat frequently, when I don't think we've changed it in 15 years.

[With MCS] we wanted to make sure we could bootstrap our customers, put skin in the game and still have the lion's share, by far, of everything we do go through partners. Those basic principles haven't changed. But the susceptibility to the nuance of our body language, and the sin of commission that somebody made in Dubuque, Iowa, when we had too much capacity and somebody bid for a job. Look, I know there are sensitivities to these issues. But the strategy, I can fairly say, hasn't changed.

Worldwide, we're just under 4,000 people roughly in MCS. That will grow this year 10 percent, 15 percent, 13 percent, 8 percent--I don't know off the top of my head, but not an order of magnitude. But how many people think we can solve the services needs of the world with about 4,000 people? We can put skin in the game in important projects. Those resources should be able to help bootstrap our partners. Yet, I'm sure, there will be a case where a partner dollar could have been bid instead of a Microsoft employee dollar. But if you look at the total number of people employed in our channel, we have 775,000 partners and 5,000 Gold Certified Partners worldwide. Do you think they all employ more than one person? My guess is yeah. Of the total service capacity of our partners that serve large enterprises, are we 1 percent, are we 2 percent [the size]? It's this order of magnitude.

CRN: We talked about Google. We ask you every year who's at the top of your threat list this year. One year you said AOL, and then you said Linux. Is Linux over already?

BALLMER: Linux never goes away. But can I say we had a pretty good year? We had a pretty good year in terms of perception, cost of ownership, security and tides turning. We had a pretty good year in terms of unit sales, share and presence. We have a great year coming with innovation with Whidbey and Yukon with [Windows Server 2003] R2.

Hey, we'll be competing against open-source products--partly because they have a different economic model--all of the time. But we've had a pretty good year, and I'm excited about it. If you ask who are our top competitors, I'd list [vendors of] open-source products. I'd certainly list Google and Yahoo. I certainly wouldn't distinguish Google from Yahoo. Both companies are doing some pretty good stuff, and it's a new space we want to be active in. If the Xbox guys were here, they'd say I might want to mention Sony. We compete and we partner with Sun [Microsystems]. That's actually working quite well, both on the partnering side and the competing side.

CRN: What kinds of companies is Microsoft possibly looking at to acquire?

BALLMER: I'm not going to answer that question. But take what we're doing in business intelligence. A huge growth opportunity. You want to talk about something. There's a lot more than we're even scratching the surface of with BI. Big ambitions. Realtime communications and collaboration. Big ambitions. Document management, life-cycle management, security. A lot going on. Management, search. Big ambitions. The issue for us isn't finding more opportunities. There are more, and we will do more. The real issue for us is making sure we're getting the pedal well down to the metal on the set of things we're already invested in.