Gates scopes out the business landscape
As Microsoft very publicly wrestles with Windows Vista, it's also quietly extending its tentacles into the business world.
Although Microsoft announced last week that Vista and Office 2007 would not land on new PCs for the holidays, the company on Sunday said it is largely on track to deliver new business software.
After a long week, and very little weekend, Microsoft Chairman Bill Gates arrived in the Lone Star state for the Convergence 2006 trade show to offer an update on the software maker's least well-known unit, Microsoft Business Solutions.
Cobbled together from a series of acquisitions that began with Great Plains Software five years ago, Microsoft has quietly grown a unit that sells the same kinds of software as SAP and Oracle, but sells it mostly to midsize companies.
In an interview with CNET News.com, Gates discussed Vista challenges, where Microsoft is headed with MBS, and the company's broader move to offer hosted software. He was joined by Doug Burgum, the former Great Plains chief who is stepping away from running MBS day to day, but remains the unit's chairman.
Q: Windows has grown over time, and one of its strengths is that it works with all the hundreds of millions of Windows PCs out there. How do you take something that complex and make it so that it can be grown in an orderly fashion and shipped on a regular schedule? That's been a problem with Vista.
Gates: Architecture layering is a key part of that, and that's something we've put a huge investment into (for) Windows this time around. You know, even in the last year, as you've seen a lot of releases of Windows Media Player or Media Center or Tablet PC, we've been doing a lot of releases. The biggest one we did, of course, was the security release, XP SP 2, and if there's anything where the amount of work you see at the top versus the amount of work that's gone underneath, you have the highest ratio of security work. In XP SP 2 we tried to make sure there were no user-interface changes. There were a few things in the browser where you just had to think about add-ons that you didn't have to before, but other than that we were able to make it mostly invisible. You know, we have great technology to test for compatibility and how we change things to avoid breaking compatibility.
When does it get to a point where it's just too hard to add new features, such as a new file system like WinFS, into Windows?
Gates: We do a lot of architectural work in Windows every release, and probably more this release than any other. It all comes back to (this): You've got end users and developers, and if you make things too complex for either of those audiences, then you're not doing your job. And the idea that if you say, OK, to make our engineering easier let's have the file system here and this fancy file system here, but the user has to think, OK, I have these concepts up here, I can do rich query, standing query, notification query up here, down here I can't, I have limited properties and things; if you force them to learn both paradigms, it's bad. And so what we've been able to do in Vista is take the search enhanced file system and get a lot out of that. It's not the database-driven file system (WinFS) that we had in mind initially, but you get extremely high percentages of the things we had in mind by using the search-based file system.
Microsoft did a lot of work to combine the consumer and business versions of Windows into one code base with Windows XP, and most people find it a lot more stable. Some folks, after last week's decision to delay Vista, were saying maybe Microsoft should go back to having more releases for consumers and fewer for businesses. Do you have a sense of what it is that people want in the next version of Windows? Is it different for consumers than for businesses?
Gates: Businesses often move in waves where they'll upgrade many things at the same time: their own applications, Office, Windows. They like to roll things out in groups so that the business processes or user training, the support gets aligned around a whole stack of software. And so you have businesses that are very quick to get everything out and you'll have businesses that tend to lag in getting things out, and then you'll have other businesses, just because of the rhythm, they'll hit our cycles when they want to make changes or they'll be off our cycle. So sometimes they'll be very state-of-the-art and sometimes they'll be a few years behind.
Security issues have made it more imperative to get up on the latest technology, and so that's really meant us making it simpler to test what pieces you have in your environment, how easy it's going to be to make that transition. There's a lot we're doing to make the transition easier, and there's more we can do to really make it rote for somebody to say, OK, I have this in my installed base, let's see which of those Microsoft has already tested, let's see which of those are unique to us--let's have things to automate the testing that we feel we need to do for those things that are unique to us.
Does it make sense to have Windows be something that is updated like Office on a pretty regular basis, regardless of the level of innovation that's there? Should there be a new version of Windows every 18 months?
Gates: (Features) like the browser user interface, the media capability, some of those things, you can have updates more often than even every 18 months, and users who want that can download those things, because they don't affect compatibility. Whereas the file system or the scheduler, the rights protection pieces, the device driver interfaces--those you're never going to modify more often than every three years or in many of those cases you want to leave those things alone for way longer than that.
Take device driver interfaces. You might let there be additional APIs, but you're going to still need to run most of the drivers that were written 10 years ago. So layering is the key here, and consumers may upgrade some of these things like the browser more often than businesses. That's hard to characterize. The one nice thing we've seen with consumers is they really use Auto Update, and so we're sort of their IT department in terms of updating. It's more complex for businesses, but even there the progress over the last four years of getting SMS (Systems Management Server) to be deployed and used and understood--how they take the updates in and when they pass them along to their systems--we've done super well on that, but it's not as simple as somebody just choosing to have Auto Update come from us.
Shifting to the business software world and Microsoft Business Solutions, it's been about five years since the acquisition of Great Plains. How has the business software landscape changed since then?
Gates: Five years ago we saw both a great business opportunity and a chance to improve the other pieces of software. So we feel great about the architectural insight and the business insights that led us to get together and really build one of the big applications businesses. We do think (that) over time there are things that--as we gain more and more momentum--the richness of add-ons that you get and the ability to invest in R&D, the ability to leverage these other pieces--we see it as a big growth business for us going ahead. In fact, if you took all our different businesses, we'd say we have more headroom and growth in MBS than almost any of our businesses. Some of our businesses in the consumer area where we're kind of new to them--then you can get radical growth rates. But of the things that are already pretty large, MBS still has so much more that it can do. We're super happy with the growth we've had the last year--it's about 17 percent.
Originally, the goal was to get to a single code base underneath your business software. Is that still the goal? What parts of the software need to be common and what parts don't?
Gates: At the last Convergence we really clarified the road map of how we would be evolving the code bases to have more commonality over time--not setting the top priority on how much code is actually being shared, but rather on the continuity of (the) user interface, as we're publishing the Web Service APIs, knowing what the commitment to compatibility for those things are. So the big expense in this industry is the cost of training developers to do add-ons, the work they do in those add-ons, and then people learning the user interface and the business processes that come out of those. Having a road map about how those things can stay the same is important for people as they invest far more than how many lines of code we have.
In terms of actually merging the code together, it's a reasonably conservative strategy because of this desire to preserve these other pieces and get moving very quickly so that all our customers have the new UI, which includes the role-based capability. The UI keeps...moving forward. Office and SharePoint have moved their UI forward a bit. So (the concern was) that there would be some big discontinuity. I think we've gotten past that, by and large. Certainly for the big customer bases--Axapta, Great Plains, Navision--they understand exactly how we're working with the code base.
Burgum: One of the things that Bill's worked closely with our teams on, too, is really thinking through how when we start addressing the commonality of the core business logic, of how we express that in a model-driven way back to the UI. We see (that) coming in the 2008 time frame, but we think that that's another big transformation. As Bill said, we're going to do that in a way where we preserve the business model for partners, preserve the investment to customers and get there in a very smart way.
Microsoft has always characterized its target market for business software as medium-size companies or departments with companies, not the Fortune 500. Has any of that thinking shifted as there's been consolidation in the industry? Oracle has been grabbing lots and lots of companies. Is it still realistic to have software that really targets the midmarket and doesn't target small businesses or large enterprises?
Gates: (Our) software can scale up to cover a super, super high percentage of all businesses in the world. When (companies) want to pick a new software application base, we will be in there competing in 95 percent of the cases. Now...there will be a small number of cases where we'll say, no, our solution doesn't have the maturity for that, we won't be pushing for that. But it's very hard to characterize. And once somebody picks one of these applications, they tend to stay with it for a five-or-more-year period. And then we're also competing to say, OK, given that you have that, let's surround that with productivity software and the other things as well. And so we'll never have this simple formula that says who we go after with our application software. And as we see opportunity, we have some of these code bases that really are scaling up to some pretty big customers.
Burgum: Our ability to scale also rides on the other great investments that are happening across Microsoft. There is a bit of a parallel with SQL Server, where there was maybe a spot where SQL was and now you see Small Business Editions of SQL doing really well, but you also see SQL scaling up into the largest of enterprises, and I think there's an opportunity for us to trail that.
Gates: Yeah, the truth is, the word "scale" is a little confusing here. The big issue nowadays isn't performance; I mean, we perform super, super, super well. It's within any vertical, how complex have you made the descriptions of products: How many tables do you have to describe that? Within any industry there might be one of the ERP packages--SAP generally more than the others--(that's) gone into more depth for somebody who wants that kind of complexity.
One of the big changes in the industry since Microsoft bought Great Plains has been software as a service, the Salesforce.com approach. Microsoft has talked about that as a capability you need to have, but not necessarily as the endgame. Is that still how you are looking at software as a service for this part of the business?
Gates: Everything Microsoft does, over time, will be available either running as a server or you can run it on-premise or it can be hosted. For most of our Business Solutions things, we have partners who are doing some of that hosting today. There are things that we're doing in our software to make that hosting work easier and easier, separating out the idea of how you administer when you're just running the hardware resource pool, versus (when) you're the business and you want to set certain parameters about how you can access what information. We see a lot of demand for on-premise. We'll certainly meet the needs for people who want to host this stuff. We think that's a perfectly valid model. Salesforce.com in a sense has gone with a very expensive sales approach, along with hosting. Most people who do hosting say, OK, it's here and they go with fairly modest sales and marketing investment, and so they are kind of an unusual combination.
Gates: Well, for most companies, if you think of their IT budget, a lot of the money goes (to) personnel costs today, where you're managing systems server by server. A lot of the automation of making operations simpler can be done for on-premise or off-premise. And the main reason hosted will make sense is where you just want to get something up and running very quickly. If you don't think you can get the IT expertise within your business, then the off-premise (version) may be attractive. The things that people still have to feel very good about is that they still have control over how quickly things get done on their behalf, how their information is maintained. If there's not enough resources there that they get to choose what can and can't run, they still have that administrative control that they can do integration.
The industry is going through that evolution. E-mail and Web sites have been the easiest things to say, OK, I understand my administrative boundaries and my integration boundaries for e-mail and Web sites. Of the stuff that's hosted today, those would be over 90 percent of it. Instead of buying a copy of SQL Server, I just want to open up and do a set of databases in the cloud. We'll get to that point where that's a very typical thing, particularly if you just want to try something out, you know, get it up and running quickly and then you might shift back later. The ideal for the industry is going to be if we have one architecture that (allows) you to shift things back and forth between on-premise and off-premise very easily, because it doesn't take much in terms of these various factors to make you want to switch in one direction or the other.
Are services like Windows Live and Office Live going to become more oriented towards midsize and larger businesses, or are those same types of approaches not as germane to larger organizations?
Gates: I do think there will be a tendency to have more on-premise with larger businesses than there are with smaller businesses. And consumers are sort of the extreme, but even there you'll still have servers in the house that are holding videos and music. Now, the administrative model of, if there's an error that comes up on that server, is it somebody in that house who has to look at that error message, or is there somebody they've got a relationship with who can look at it and diagnose it for them? That's kind of like hosted, when you're able to remote all the management of the thing. And in the history of software, people have been doing forms for that remote management stuff, you know, going back to before Microsoft even existed. Now, as we're making the software hostable, we're also saying even when you run it on-premise, if you give me permission and we have the right relationship, I can be examining the health of the thing and helping deal with certain types of issues that come up, even though it's your hardware, it's on-premise.
Some of that remote management technology is in Vista, right?
Gates: Remote assistance is (in) Windows XP, but it didn't work well through firewalls, and so now we've improved that and we think that will be a lot more mainstream than it is today. Most tech support calls today are still voice, you know, picking up the phone and describing what's on a screen to somebody at the end of the phone line. It should be that you're sharing that screen or you just describe the problem and that remote person takes over. So we're seeing that evolution of better software capability for letting the remote expertise be applied, no matter where the code is running.
