Microsoft Internet Brokerage and Banking Conference
Remarks by Bill Gates
Meydenbauer Center
Wednesday, March 26, 1997
Bellevue, WA

BILL GATES:

Well, good morning. The last year has seen a lot of progress and a lot of change, and that's true in every aspect of the platform we'll be talking about over the next few days.

First let's talk about the PC itself. Every year we get the incredible benefit of Moore's Law, which says that because of the power of chip fabrication, every two years the machine will be twice as fast. Now, we can take that amazing power and use it in two different ways. One is to deliver machines with more capability and the other is to deliver machines with lower prices. And with exponential rates of improvement, you can actually do both simultaneously.

And so, we're starting to see very reasonable PCs, including displays, being sold for under $1,000, and these are machines with over a gigabyte of storage, 133-megahertz Pentium processor, graphics, CD-ROM, not much of a compromise for a home machine. In fact, they continue to include the most up-to-date modem capability, 33.3K baud modems, to get connected out to the Internet.

The structure of the PC industry is a very important element of how innovation gets out to users very rapidly. Innovations in hardware and software are separated. Because we have Windows as the center there defining the standard, customers have a choice of how they're going to buy the fastest machine, the most portable machine, and they can change those choices or intermix them independent of what they're doing in software.

In the PC space, you know that you're going to get the most choice of applications, the most choice of peripherals, and because of the volume, all of those things will be far less expensive than they were in other computing environments. For example, the Unix environment uses the term "open" a great deal, but the flexibility for the customer after choosing a system where you get locked into the different security and user interface, it's very, very different than the world of the PC.

So this structure takes openness to a new level, and as the PC has come down in price, as the volumes come up, the richness of the software has simply tracked with that.

Now, the big developments in the last year here is not just what's happened with PC volume going to over 80 million, but what's happened with PC servers. A few years ago, you could say, "Well, PCs are great desktop machines, but for high-powered applications that used to run on mainframes, you need to buy more expensive equipment." But with the improvements in the Intel chips, with the improvements in Windows NT, there's been a huge shift now to move even those server applications onto PC architecture.

Part of this is the multiprocessor machines that have come out. It's very typical now to get an off-the-shelf machine with two or four processors, and in the next couple of years, it will be very typical to get a machine with eight or 16 processors. So you have all these elements -- individual chip speed, using multiple chips, and improvements in the software -- driving performance up.

To give an example, on these standard industry database benchmarks, TPC/C benchmarks, we've gone from 600 transactions a minute two years ago to now 6,000, and we're confident that by throwing clustering into the mix we'll be at 60,000 within the next two years. So that means now we're surpassing even very expensive platforms in terms of the capability that's there. And so the simplicity this can provide, as you get a more homogeneous environment -- common security model, user interface model, API model -- between the desktop or portable machine and the server, it's really quite dramatic.

The key element that Microsoft provides for this is Windows NT. We made a huge investment and took a huge risk in building Windows NT five years ago, and there were a lot of people very skeptical about what would happen with this. Would it get the critical mass that was necessary? The design was to take the performance of NetWare in terms of file sharing, the richness of Unix, and the user interface of Windows and put those together as a reliable, secure, and scalable platform.

Now, over the last five years, we've built a momentum here, whether it's tools or pioneering customers or vertical applications. There have been some very, very major wins. People like SAP, who used to be focused on the mainframe, now have Windows NT as by far their most powerful platform. And so that's a case where one of the most demanding applications in the world certainly runs by far the best on Windows NT. We run Microsoft totally by using SAP running on top of Microsoft SQL Server and Windows NT.

Now, the volume of Windows NT is a key thing. Again, you've got that success loop where the more volume you get, the more applications you get. The more applications you get, the more volume you get. The Unix world never had that because there was total fragmentation. Even though people would go to committee meetings and say, "Okay, let's agree on this or that," their engineers were always working on things that were different between the versions, and so porting and testing on different instruction sets, and the different software base, always made it very difficult.

In this chart, I've taken all the flavors of Unix, over 14, and aggregated those together. If you compare server volume, in calendar '96, NT Server moved up by about a factor of two while the other platforms, NetWare 4.0 and Unix, stayed at about the same sales level. We do a lot to interoperate with both of those systems. There's a significant installed base out there in those sales, even though they'll decline, it will be important for us to work together with those systems. But the momentum is very, very strong.

At the bottom, I show a particular statistic about Web servers. A year ago, Windows NT did not have Web server capability. It was just about now ten months ago where we ship with our HTTP server, which we call Internet Information Server, bundled in with NT, and we were able to demonstrate that our Web performance is better than the most expensive system you can buy. Even an off-the-shelf Pentium-type system can give you these incredible benchmarks.

And so, in less than the course of a year, Windows NT has become a majority of the servers out there on the Internet. Again, there are a lot of different operating systems that are used. We just passed over, in the fourth quarter of '96, 51 percent for Internet and 64 percent for Intranet. So a lot of momentum for NT in the general marketplace, and specifically with financial institutions, we've done very, very well.

The design cycles, of course, only come up every four or five years for things like branch banking systems or broker systems or trading-type systems, but as those have come up over the last couple of years, we have won the vast majority of those. We're in over 350 banks, which is 120,000 branches now. That's the actual deployment, and the planned deployments would take us to almost double that number.

We are in a lot of automatic teller machines, including at CitiBank. A lot of big wins in the call-center area. In brokerage terminals, we've done very, very well. Six of the top ten firms that have gone to new desktop devices have chosen NT. And then with Web servers, I showed a few examples there of people who drive their Web site off of NT technology.

The financial area has been a particular focus for us because the customers there are very demanding. The opportunities, when you can deliver better software building blocks, are really quite incredible. And so, we've had a vertical focus with lots of specialized people calling on customers and understanding the needs there and setting standards like the financial connectivity standards, and then, of course, events like this.

The key message about NT is not just to use it for a Web server or just to use it for a branch banking system or just to use it for a call center. Each one of those things NT can win on its own merits, but the biggest benefit of all is if a financial institution thinks of its applications as delivering to all those channels off of one development effort. And so the information that somebody at the call center sees on the screen is a superset of what the customer sees when they come in through the Internet, and they see the same thing when a teller is looking up that customer or when somebody is trying to analyze the customer behavior.

And so sharing across those delivery channels allows you to move to have all the information about the customer in one place and to have it presented in a way that's very, very easy to understand. And NT can play a key role in that alignment. In fact, there's many ways to get there. You don't have to replace the back-end database systems. You can simply tap into those systems with standards like ODBC and present the information, although it comes from different places.

In the long run, the ideal approach is to have a unified database, but for many institutions that is on a different time scale than just getting the presentation correct and making it available to all the different channels.

Now, a lot of what we're going to focus on in these two days is how the Internet is allowing customers to get at information in exciting new ways. The Internet is a fantastic development for financial institutions. It will allow you to understand more about your customer behavior. It will allow you to serve the customer in a more customized fashion than ever before. And there's nobody between you and the customer.

What you put out on that Web site is yours, and it doesn't matter what technology you use. Every bit of how you present the information is completely under your control. Of course, the customer has to pay some communications company to get connected up, but that's very neutral. That's just a pipe that lets them get connected.

The Internet is such a phenomenon that certainly people have gone overboard in some cases, saying that everybody will be using it in four or five years. That's not the case, but every year the Internet will grow in popularity. The variety of content, the quality of the content, the speed at which people access that content, the security elements that let you know that you're the only one who can get at your information, all of these things are improving very rapidly.

The level of investment of entrepreneurial firms in this area is almost unbelievable, and despite the fact that some of those companies won't succeed, many of them will bring lots of great things into the Internet arena. So any weakness you can name, Microsoft and many other companies are working on eliminating those things.

And certainly we're at critical mass, where it's just going to grow and grow. Electronic mail, perhaps one of the more mundane elements of the Internet, is very important. We think that not just knowledge workers, but eventually people broadly, even out of their home, will find electronic mail very attractive.

I mentioned the speed of connection, and that's one of the trickier issues. People in business or at universities have LAN connections that give them very high speeds, but out of the home, for the next decade, the vast majority of people will be using a dial-up phone line where the modem speeds will go up to 56KB and that's it. The infrastructure simply doesn't allow you to go faster. It's not too bad. It lets you do still images pretty well and certainly text is very fast.

But the most engaging experience comes as people move up to ISDN or ADSL or cable modems, but the cost in rolling those out are high enough that it will be a minority of people in the years to come. Eventually we'll have two-way digital video. That dream, that broad-band dream, is the original interactive TV dream, and it's something that the industry was very excited about three or four years ago but was quite unrealistic about the cost of rolling it out and what the revenue picture looked like for it.

And so we'll get there, but the time frame is a lot longer than people expected, and evolving based on Internet standards is a key part of it. At this point, we have more people involved in Internet-related committees than any other company, and the standards are central to how all of this will work. In fact, I'll be talking about some of those new standards later on.

When we think about the Internet, PCs are 99 percent of what's hooked up today. In fact, it's interesting that only about 25 percent of all PCs are hooked up. The majority of people are not hooked up. You might say why or ask why. Well, paying $20 a month actually is a lot more expensive than buying a PC. A $1,000 PC not connected to the Internet is cheaper than somebody giving you an Internet device and saying, "By the way, you have to pay $20 a month and tie up your phone line to go and get things."

And the response times and richness, if you compare, say, the multimedia encyclopedia that you get with most of those home PCs, you compare accessing that information off the CD, it's about 50 times faster than going at that same information across the phone line. And so we still have a lot of work to do to get a high percentage of PC users connected up.

Now, over time we will have little hand-held devices that'll connect through wireless connections. I know we have a session talking about how we're using Windows CE and working with hardware companies to drive that market. We will have intelligent TVs where you use the cable infrastructure and a rich set-top box to connect up.

The device that you edit text on and that you sit close to, though, we'll continue to call the PC. And the PC will evolve far more rapidly than Internet speeds will. That is, in four or five years we'll be using voice input. We'll have flat-screen technology. And the kind of graphics, motion video support in the PC, will be pretty unbelievable. The only thing that'll hold things back is the speed of that pipe that gets you to the Internet. It will continue to be a limiting factor.

Now, when people use the Internet, they don't just go there because of one application. They don't say, "Okay, I'll sign up to the Internet just to do banking or I'll sign up to the Internet just to buy books at 'amazon.com'." You basically get comfortable with it as a tool and then use it for all your information needs. You go and get sports scores. You see the latest news. Before you go and buy a car, you find out what the dealer paid for it so you're a little smarter in the negotiation, because that information is easily available. And so there's a wide range of ways that it gets used.

In the Internet case, you're mostly reading information. You're not creating information. That's a big difference between Intranet, where the primary tools, Microsoft Office, that are used for creating information need to be part of the picture, because you want to view data in different ways; you want to annotate and create it. The knowledge worker is not just a reader, whereas the Internet browser is primarily a reader, and so you get a little bit of difference in terms of what software elements are critical there.

I think that over time, paper forms will disappear altogether because it is so much more efficient to have people fill things out electronically. Certainly at Microsoft -- and I had a copy of every form we use in the company delivered to me and then I went through with each group that had those forms and asked what was the plan to get that online. We've already gotten rid of over half of them, many of them related to human resources, things like 401(k) plans or stock plans. Before, it was so cumbersome to get your status and check things, make sure you could do something. Now that's a trivial thing to do online, and people just have a link on their desktop they click and they get to all of these HR-related forms are just there up on our Intranet.

Sharing work is a big thing across the Internet. We have a thing called Net Meeting that lets you share a screen with any other Internet user. That kind of collaboration, we think, is very important. And transactions will be a key element of what people are doing.

Now, with this electronic medium, there's a lot of excitement and a lot of fear about what it means. I think that in a way the Internet just magnifies the excellence of businesses today. If they're good at tracking their customers and tuning their products to customers, this will allow them to do that in a better way. If they're not good at those things, just the fact that the Internet is there isn't really going to change that. Competitors who have those skills will be doing a lot better.

There are opportunities to reduce costs, to streamline things. More importantly, the kind of quality of service, the ability to get around geographic boundaries in terms of reaching out to suppliers and customers, those are very substantial. The early applications are substitution, where you take paper forms and you make things electronic. You take information that people have a hard time getting, like the budgeting information, and you make that a few clicks away, completely self-explanatory with mail that you can send out if you have any questions. The more radical applications have to do with changing the boundaries between you and customers and you and suppliers.

Now, all of this is happening pretty rapidly, and so the interesting thing is that these tools, these information tools, the PC connected up to a rich network which in turn is connected up to the Internet, they are very critical for business to move it at the speed that the marketplace requires.

A lot of this has to do with how you can get more out of your knowledge worker. The basic entry fee is to have the network, have electronic mail, have the productivity tools there. But once you've paid that price, then the question is how much value do you get out of that. And there'll be wide variation from company to company of how that goes.

The great thing about what we're seeing in this conference is that the things you want to do to get your internal knowledge workers working more effectively are exactly the same things you want to do in terms of having an Internet site where you can directly work with that customer, the same kind of easy information access. So it requires re-engineering for new capabilities.

Microsoft's role is as a technology provider. We have some amazing building blocks, starting with Windows NT and BackOffice and Office that make it very easy to build these systems. Another key message out of the conference this year is that the amount of time to build a very effective transactional Web site can be measured in months. We're putting together this tool kit called Marvel that you'll see more about. It takes a lot of the steps and shows you exactly how to do those. But building your own Internet site is not that hard. And it's such a critical thing that you really will want to think through how the information is presented there.

Some of the general initiatives that Microsoft has include the focus on what was called Total Cost of Ownership. It's really not the right name. It's more a partial cost of ownership, because when you empower knowledge workers, you not only have the desktop costs, but you have the networks and the servers and the communications charges and application development charges.

In many of the plans people look at here, they eliminate all the power and responsiveness on the desktop, and so they shift all the costs up to the servers. They overload the network and they ask you to rewrite all the applications. And so just pushing down one cost to have much greater [costs elsewhere], while getting rid of the flexibility -- that's not a good solution.

What is a good solution is to take what's been hard about PCs -- software distribution, keeping all software updated identically -- and make that automatic, build that into Windows NT. And so we have a wonderful solution where all the state of the PC actually gets mirrored up on the server, but we make sure that because we have a local disk, 99 percent of the time you're accessing it locally.

So you get the speed of local access but you get the richness of having the data centralized so you can go to a new PC and immediately you log in. It knows what your files are, what you're interested in, and it's just available to you. If you want to go to a neighbor's machine, replace your machine -- all very automatic. And that's why we say "zero administration" at the desktop level, because the desktop just takes all that central state and brings it down so that it'll be very rapid to get at.

We have some big initiatives in terms of making it easy to write server applications, building transaction management into the operating system. And all these rich new things we're doing work across all languages -- Java, C, COBOL, whatever language people have used -- so you can write new code, you can mix languages together, but you can also preserve the code that's there.

There's a lot we're doing with integration. To simplify these systems, you can't think that administering your data bases means something different than administering your file system, which is different than administering your mail system. That's just too complicated. You've got to have one directory that manages everything, one security approach driven off of that directory, one set of storage APIs that affect all those systems. And Windows NT is the place where all of that comes together.

So Microsoft has a lot of pieces. We center those pieces around Windows on the desktop, Windows NT on the server, Office and BackOffice. We're also doing some new interactive things on the Internet -- Slate Magazine, MSN, our news joint venture with MSNBC. We're getting very good volumes there. Over a million people a month visit or Web site. We have two and a half million people on MSN. Those experiences are teaching us a lot about these tools and how to make things compelling. And we will have some new brands there. None of those relate to doing financial transactions. That's an area where we've made it very clear again and again that we will not be directly involved. We are an enabler.

One of the key new products is the update of Internet Explorer 4.0. Brad Chase will talk about it a little bit more this afternoon. But it drives a lot of the things going on: the integration of the views, the bringing together of sending information to you as well as you going to get information -- so-called push versus pull -- and richer collaboration of communications, things like the conferencing, the Web phone capability, built-in, very easy to use, as well as being the best browser -- faster, better user interface, and, of course, free on Mac, Unix, 16-bit Windows and 32-bit Windows as well.

So just to reiterate, we see ourselves as the supplier to you of the core systems, the core development tools, all of these now over the last year, putting Internet standards at their core. We work in a very leveraged fashion. We work through solution providers, many of whom are here in the audience, that are more specialized in terms of financial activities. We do think that our consumer finance software, Microsoft Money, is a very handy thing for people who have PCs to keep track of what they've done and to monitor what's going on. And now, with the new standards, we've made it very easy to hook that up to your transactional Web site.

The key new standard here is OFX that we'll talk about quite a bit more. Making it easy to build the transaction platform; that's what this Marvel product is all about. Shared development: that's a theme I've talked about here, and we'll make that more concrete over the course of the next two days. Richer customer experience: that means that when they come back to the Web site, they don't just see the same thing as everybody else. They see an offer and a status that relates to them in particular. And anybody who's ever wanted to know what customers think, what customers are doing, they will love this environment.

And there's a lot of tools we've got to help you track what's going on in your Web site, not just simple things like what pages are most popular, but looking at customer characteristics and see where they visit, which links they follow. And you can even put logic in the server to try different offers randomly to different people and then see what those responses are. And literally within a 24-hour period, you can have that feedback.

I'd say the bottom line here is these are very early days in the world of the Internet, but the opportunity to be a pioneer, to get the young customers who really are seeing the Internet as something they'll go to every day and get information, that opportunity is pretty incredible. And we're certainly very excited about pioneering that together with you.

Thank you.(Applause.)

Q&A

LEWIS LEVIN, Microsoft vice president of Desktop Finance: We have time for a couple of questions.

Q. Good morning. I'm Jennifer Bloom (sp) from the American Banker. I learned this morning that IBM, Netscape, Sun Microsystems and Oracle are planning a press conference this afternoon to announce standards for smart cards in network computing. And my question is twofold. One, what's your take on the future of network computing? And two, does Microsoft have plans to deploy smart cards?

MR. GATES: We do have a lot of people using smart cards together with PCs. For example, if you want to replace a normal log-on where you just do a password with the smart card authentication, lots of corporations, particularly banks, are doing that today. As smart cards roll out and get more popular, we see that smart card reader as being built into the PC keyboard. There is a standard for what that physical form factor looks like.

The smart cards we like are the ones that have public key capability, and the price of those is still a little bit high, but over the next couple of years that will come down quite a bit.

Remember, NCs are sort of a two-letter form of Unix; that is, the Sun NC is very different than the IBM NC is different than the Oracle NC. They don't like the Windows world that makes hardware so competitive. It's been a real problem for their margins. Companies like Compaq have adjusted to that environment and have the high volume to make it work very well. But Sun has always lived in this high-priced world, so they want to control their own system software, so their engineers do something unique for the Sun NC.

People talk about the NC as something thin, but you have to remember that the Java run-time is an operating system, and everything that's in Windows today is being built in there, as well as a browser. The recent Netscape browser just barely runs in a 16-megabyte system, and it's the fastest-changing piece of software there is. So anything that runs a browser is not thin. It's as thick as a PC is. In fact, it's sort of outgrowing the PC. Certainly, it outgrew the install base. These latest browsers will never run on Win 3.1 machines, because those machines just don't have enough memory to grab those browsers.

Anyway, it's a topic I could talk about for a long time, but you should feel pretty safe with the PC in terms of the software, the price, the availability. We really have to prove something to customers in terms of how we're going to simplify things. I mean, the only reason people dream about this illusion that the NC would help is that we haven't done as much to make it easy to administer PCs as we should have. And, in the same way we surprised people on the Internet in the last year, this year what they'll be even more surprised about is how we can simplify administration, including things like moving all the state to the server like I talked about.

MR. LEVIN: Another question?

Q. Hi, Bill. David Wiseman (sp) from Forrester Research. You talked a lot about the new Internet technologies. I was wondering how you saw the emergence of an Internet desktop, push technologies, component software, and collaborative technologies changing personal financial management software like Money and Quicken. What is your vision of how the personal financial management software will look in the year 2000?

MR. GATES: There are a lot of elements there. Information browsing, you can have information sent to you and show up in an in-box, or you can go out and get it, which is you type the URL in or you follow a link and go get that information. That's push versus pull. And both of those are extremely valid. In fact, they really go together. Where sometimes somebody pushes to you a notification that something has changed and they include some links in there, you can either go follow that immediately or you can just save that away and go and get it later.

There were about 20 companies that came up with standards for the idea of pushing Internet information. And we said, look, this is the world of the Internet. We should have an open standard here. You know, browsers are free, and so let's get the W3C and other committees working together on this. And so we've proposed what we call a channel format that allows you to do push and every browser will just have that built in. And so with push, all information publishers will not only have their site you can go to, but they'll let you sign up to say, let me know when things change.

Some other examples of that for this group, if I belong to a mutual fund, I may say, you know, send me mail when you have a new offering of a certain type, or let me know if my net worth goes down by 10 percent, I may change my purchasing behavior a little bit.

And so you so you'll set criteria for stock suggestions or developments in the marketplace that you'd like to be notified of, and all of that will work within the browser environment. You won't have to have special software. You'll continue to have Microsoft Money and Quicken -- for somebody who wants to really track their transactions from multiple sources and look at their budget and tie that into electronic tax filing.

MR. LEVIN: Yeah, I think there are two distinctions that are emerging in some of the push technologies. I think the early versions of push involved publisher-driven deals deciding what to push. By going towards standards on push technology, we're going to return the Internet to its roots, which is as a user-directed experience. And so you'll have more people making the choice in what they want to pull, which makes a lot more sense for what financial institutions do where individuals will decide what they pull from their specific institutions.

In terms of the impact this has on financial software, I think we'll see the distinction decrease between a large piece of software that you buy all at once in some sort of package to things that might come from the financial institution and that come in smaller chunks, but that will be a gradual transition because we don't want to turn customers into systems integrators. But you're going to see that happen over time. And you'll also see the personal finance software as being a much better receptacle for information that does come off the Internet rather than from some proprietary stream.

Q. I'm John Harrison. I work for National Financial Correspondent Services, so I'm actually in the customer group, not on the analyst or the publisher group. Many of us are here because we intend to or are in the process of creating OFX servers. Many of us were very heartened, I think, by the announcement of the Intuit/Microsoft Check-Free collaboration, which is OFX. That brings the promise, I think, in many ways, of openness and the ability to get to market share with one development effort, not two or three or four development efforts.

When we hear you say that, you know, the problem with NCs is that they are flavors of them, one of the concerns that I think many of us have is that there will be flavors of OFX, because if history is any indication, Microsoft will embrace and extend OFX and its implementation, and in an attempt to eliminate the Quickens of the world, and we don't necessarily want that to happen because those folks are our customers too. Can you -- is there any reassuring phrases that you can give us that would, you know, make us a little more warm and fuzzy for the next few days?

MR. GATES: Well, Lewis may want to answer this, but we initiated the OFX discussion because we thought that would grow the market. And it is an open standard. Anybody can implement OFX. We're making it particularly easy to do with some of the tools we'll talk about here. The goal is to grow the market, make it easy for banks to get up. And the percentage of banks that accept transactions on their Web site today is absolutely tiny. And the market for personal financial managers as just stand-alone products is a modest-sized market. I mean, we have our share of that business, but it will be a much bigger business. And the other elements we can bring to this, there will be much higher volume if we can make it an obvious thing for banks to set up those Web sites. And so that's the goal of OFX, and that's a goal that will endure.

One thing you should never worry about is the competitive nature of the software market. Software prices have gone down constantly, and take one example, Windows NT. For about 3 percent of the cost of a system, we deliver Windows NT. That means for less than $100, you get Windows NT. Windows NT, the amount of testing and man-years we've put into it is far greater than, say, IBM puts into MVS, and yet there's a factor of 10,000 price difference between Windows NT and MVS. And it's just because we are very volume-oriented, and we understand that only by keeping things very acceptable, very low priced, can we maintain a strong position.

MR. LEVIN: I'm going to talk a little bit about the progress that the OFX partners are making. We're now actively talking about the revisions process, the role of an advisory board, and the certification process. It's certainly our intention for people to have confidence that they can implement their side or both sides, whichever, of OFX and rely on that working. There will be the possibility for extensions because that's a necessary thing, but we don't want extensions to become the rule. There should always be an extremely large core subset which is substantially all that most people want which reliably works across everything.

Now, there are also going to have to be mechanisms where there's some communication between a client and a server to say, "Hey, I want to use this extension," and for the server to gracefully say, "Well, I don't do that," and for the client in a sensible way to tell the user, "Hmm, that provider doesn't support this particular thing." So we're working those things out.

Q. I'm Steve Beck with A.G Edwards and Sons in St. Louis. How do you see OFX changing the personal relationship that a retail investment broker has today with their client? Specifically, do you see it enhancing or weakening that relationship?

MR. LEVIN: I think the potential exists for OFX to help in that customer relationship quite a lot. What OFX provides is a structured way of reliably exchanging information with a client. What it doesn't do is actually change the relationship between clients and financial providers. Because it's a standard and because it will be implemented by a wide variety of partners, it's actually going to enable some dialogues that today are extremely difficult. I mean, a certain amount of time is expended in a meeting between a financial advisor and a client just exchanging information in a fairly clumsy way, and it means that the kind of high bandwidth thing you do, which is to understand the person's situation and give some meaningful advice, you're sort of preoccupied with getting some basic information in a very high-cost way. And so I actually think it enables that quite a lot.

It also, in some ways, lowers the bar. OFX coupled with the Internet, instead of having to rely on a very, very substantial investment in some large central communications capability, even a small financial adviser with a few hundred clients can have a very effective online capability directly to the customer. So, if anything, there's a leveling impact that I don't mean to say anything unfavorable towards our friends from Merrill Lynch who might be here, but in the olden days there was a huge difference between the individual financial adviser and Merrill Lynch. If anything, the Internet brings up the capability of the smaller guy.

Q. Dean Gittleson (sp) with Perot Systems. My question is not presentation layer, but back at the legacy layer. Mr. Gates talked about getting to legacy data and maybe moving those platforms to more maybe current database platforms. But NT now runs on Tandem, and Tandem is a bank platform. And are there investments, R&D investments, at Microsoft to start looking at middleware to connect to legacy data, because that is probably the most difficult part of moving financial data onto to the Web or on to the PC itself?

MR. GATES: Well, the biggest initiative we have in that area is the ODBC effort that connects arbitrary data sources with arbitrary data consumers, and it can reach into not only SQL structured data from Oracle or Informix or DB2, but also into more classic data stores like IMS. And that puts what people are using to get at the mainframe storage, and it's reasonably straightforward to extract the data in that way from multiple sources, and then present it in an integrated way. That's why I said there's two ways to approach this. One is to leave the back-end alone and work on the presentation. The other is more radical, which is to change your data dictionary, which eventually may get there. But a lot of people are starting leaving the back-ends in place. And certainly if you want to build a Web site in a few months, you had better use ODBC connection out to that data to do it.

MR. LEVIN: I think we have time for like two more questions.

Q. Hi. Julie Kline, Retail Delivery Systems News. I want to ask you, Mr. Gates, there's a lot of banks here today, but there's also a lot of banks who aren't here and still choose to operate these systems on proprietary platforms. What sorts of words do you have for those banks that are still "anti-Microsoft" today?

MR. GATES: Well, I don't know any banks that are anti-Microsoft. The fact is, banks have their own cycle when it makes sense to go in and redo their branch banking or call center systems. And I'm not aware of anyone who, in the last couple of years, made that decision and chose, you know, OS/2 or CTOS or some proprietary system. I mean, I even have a hard time thinking what the alternative would have been to taking advantage of PC industry building blocks.

And so we're working with banks according to their plans, how they want to mix their existing systems with Windows systems, when they want to go through those upgrades, and we're doing extremely well. I mean, given the plans we set down three years ago of where we'd be in all these sectors, we are way ahead of those, and we are a very patient company. You know, what we have are reasonably high-volume, quite low-cost building blocks that they can take advantage of, and that message is becoming clearer and clearer.

There are some banks that are extreme enough, only a few, that they use Windows NT to run everything, but most are mixing together systems of various types. Numerically, the Windows systems are far outnumbering the others.

MR. LEVIN: This will be the last one.

Q. My name is Scott Hodgson, and I'm with Platt Brothers (sp) Software. You mentioned earlier that you're not in the business of executing transactions. However, some of our clients which are brokerage firms see your "investor.msn.com" and feel threatened by the fact that you're offering portfolio management and trade, or potentially if not already, online. How do you reconcile the two?

MR. GATES: Well, we're not offering trades online. "Investor" is an information site for investors.

MR. LEVIN: What really happens there is, we don't change the relationship between individual investors and the brokerages that the choose. It's a sort of jumping off place. There's an information resource there, there's a brand of visibility of the brokerage, and the customer is quite clear, I'm going to [a broker on the list], and the list is not long. So one of the reasons we're here is to make it longer. They go to the broker of their choice, and it's quite clear that you're performing a trade with Schwab or Fidelity or PCFN or E-Trade or any number of others. So it might be in the advisory role that they say, we'd like to be supplying that content, too. But our view would be, the customers have a choice about the kind of content they use, and we're not in the trading, execution business at all.