Based in New York, Dow Jones & Company publishes a wide array of vital business and financial news and information. The firm's flagship publication, The Wall Street Journal, is the leading global newspaper of business. Dow Jones also publishes Barron's, a business and financial weekly; the Far Eastern Economic Review, a Hong Kong-based weekly magazine that provides current news and analysis on Asian business, economics, and politics; and The Wall Street Journal Classroom Edition, published monthly during the school year for U.S. high school students.

Dow Jones's position as the preeminent publisher of business and financial news and information extends well beyond the printed page. Many of the above titles are available on the Internet, including The Wall Street Journal Online (WSJ.com), the largest paid subscription site on the entire World Wide Web. Dow Jones also provides real-time news delivered electronically. The Dow Jones Newswires grew from the company's primary newswire, which has been the leading electronic provider of comprehensive business and stock market news to the securities industry for more than one hundred years.

Bill Godfrey is Dow Jones's CIO. Prior to joining the company in February 1999, Godfrey was senior vice president of IT services at insurance company Hartford Life. He was with Fleet Financial Group from 1989 to 1996, culminating with his position as senior vice president responsible for the bank's corporate cash management systems portfolio. From 1984 to 1989, Godfrey held a variety of positions supporting CIGNA's individual financial services division. Godfrey worked as a systems engineer for Electronic Data Systems from 1980 to 1984. He received a bachelor's degree from the University of Massachusetts, where he majored in finance.

As Dow Jones's CIO, Godfrey is constantly looking for ways to improve the company's information distribution efficiency and to cut costs. He feels that Web services fit the bill on both counts. Dow Jones recently began using Web services to automate the creation of the published financial statistics that are a company hallmark.

In our market data area we use XML and Web services to automate a significant part of the tabular material that ultimately makes its way into our newspapers. The "C" section of The Wall Street Journal is virtually made up of scores of different categories of market data—equities, fixed income, currencies, charts, and analytics. We get data from the world's major financial exchanges. We also buy data from well over one hundred different sources—the Associated Press, Morningstar, Lipper, and so on. We bring that data into a content acquisition platform, a mechanism that takes in all the feeds and essentially puts them into a common format in a DJ-specific XML vocabulary. Web services is the glue that ties together approximately fifteen systems, written in different languages and on various platforms, that are necessary to create the different stats packages destined for the company's many publications.

Godfrey views Web services as an excellent way to boost efficiency and cut costs.

Like a lot of companies, we've been going through excruciating cost pressures these last couple of years. At the same time, paradoxically, we've been making major investments in a business process redesign and increasing productivity. Over the past year, there's been quite a bit going on in The Wall Street Journal. We had a project called "Today's Journal." This initiative created the paper's new look and feel, including color on all the section fronts, a lot more color advertising capacity, the new Personal Journal section, and a redesigned market data section. If you look at the "C" section and see how that data is presented today versus a year and a half ago, it's quite different. It's the same with Barron's—a new look and feel; a new design.

Our strategy for keeping up with the demands of the new product designs was to automate the creation of almost all of our stats. Because we're a media company, all of our products are created every day, and they're consumed every day. It's what we call our "daily miracle." So everything's got a deadline, and we have to be able to do all this and meet the internal deadlines for all our products. The stat manipulation process is complicated and time-consuming. It requires a lot of people, and if we didn't automate it, we predicted we'd have deadline problems.

Another important fact is that we're Dow Jones, which means we take quality and customer service incredibly seriously. Operationally, our processes are set up to double- and sometimes triple-check information to ensure accuracy and quality because that's part of our brand promise. So, in the market data world, people worldwide depend on The Wall Street Journal to be the authoritative source for market data. So we've had to be very careful as we automate that we protect the quality of the final product. I think we're off to a very good start on this.

Dow Jones decided to go with Microsoft technology.

We're using Microsoft BizTalk as the Web services engine. A market data Extensible Markup Language taxonomy, leveraging SOAP, forms the application strategy that connects a bunch of product platforms and creates the final form market data and analytic packages that go into The Wall Street Journal and Barron's.

Godfrey was attracted to BizTalk by its price and capabilities.

It's a very well-priced platform. Also, my team felt that the administrative capabilities of the system, in terms of writing business rules and its work flow, were simple to understand and robust enough. And its basic XML management functionality was exactly what we wanted.

Prior to adopting Web services, Godfrey and his IT team brainstormed their way through an array of potential applications.

We prototyped for about a year, and created thirty or forty prototype Web services to make sure they could provide us with the capabilities we thought they could. So, we created things like a real-time quote service, a charting service, a "publish a piece of content" service, and a "publish a home page" service. We created a variety of prototypes so that we could determine if, in fact, Web services would live up to its billing.

Most of the applications worked as expected, says Godfrey.

We've got a half a dozen of these applets in place, but we were basically just dabbling, doing proof of concept work. The first real application was our market data automation project. That was number one. And that's principally where we use Web services today—in production.

Next on Dow Jones's Web services agenda is the automation of its Newswires, which transmit financial data and news to clients worldwide.

We're in the process of redefining one of our core businesses: the real-time Dow Jones newswires. They're our most venerable business—110 years ago the information was provided on a ticker and printed on ticker tape.

The Dow Jones newswires is a global electronic business that creates 5,000 to 10,000 stories every day, and pushes them out across a variety of networks. They go to the buy and sell sides of the professional finance community worldwide. There's domestic and international equity data, fixed income data, and capital markets data, as well as news stories on energy, commodities, and other topics. Basically, the Dow Jones newswires is a real-time financial services product that's provided to professional financial markets. It's news that moves markets.

Yet Dow Jones needs help in delivering its content to end users.

This is a business where the vast majority of our content actually reaches our customers through third parties. So we push our content through Reuters, Bloomberg, and Thomson Financial, for example. They, in turn, provide terminals on our customers' desktops. These customers can be brokers, dealers, wealth managers, fixed income bond traders, and so on. We're in a marketplace that's consolidating into a few big players—Reuters, Bloomberg, Thomson, and a few other big houses. Yet the data has to reach hundreds of thousands of customers. One way to do that efficiently is to use Web services.

For content delivery, Web services provide a new level of convenience and efficiency, says Godfrey.

The aim is to create technology capabilities that will allow us to integrate our news into work flows of our customers, while lowering the overall cost of product implementations and customer service administration. Traditionally, whenever a change was required, you would have to go onto your customers' premises, or work with them, have a programmer involved. That would be true every time you rolled out a new file format, or a new news feed. So, both for the customer and for Dow Jones, the traditional model required lots of business process integration, implementation, support, and coordination every time you wanted to make a change. We can avoid all that with Web services.

The first phase of Dow Jones's "Newswire of the Future" project is a service called Dow Jones NewsPlus. NewsPlus is also a new technology direction for Dow Jones. The first phase is a new Web site, and the XML-based feed is designed to give a convenient and interactive interface to the Newswires. Dow Jones's editors select and rank the most critical, market-moving stories for Dow Jones NewsPlus, with summaries on the home page and links to the complete story and related information such as press releases, industry news, and fundamental company data. The home page also provides stock and bond market updates, as well direct links to critical topics like politics and economic indicators. The interface also highlights select features, such as columns and analysts' comments.

It's designed to take what has traditionally been exclusively a proprietary, text-based, real-time, exchange format and move it to open industry standards. The way to do that is to evolve our news delivery using industry-standard Web techniques. That means embedding links within our content, so people can click through a story to get greater insight and analytics. It's designed to include analytics and to have more fielded data and richer meta data. It also offers a lot more personalized entitlement capabilities so that we can, on a finer level, give our downstream customers precisely what they want, allowing them to mix and match content. We want to be able to provide a delivery metaphor that allows customers to choose what content they want, and then how they want to display it. To do all of these things, we had to reinvent the way in which we distribute our content. Therefore, we're making some investments in emerging Web technologies, including Web-centric editorial platforms, XML taxonomies, and browser-based navigation.

What it comes down to is that we all want to personalize the desktop. And, if we want to tailor the data, the content, or the news to the specific requirements of that individual professional, then we need technologies that will enable that. The browser alone isn't enough. A fat client GUI can't do it. An old sort of model on a mainframe screen also can't do it. The idea in our business, in terms of real-time newswires, is to integrate news into the work flows of our customers. To do that, you need more than HTML. It just won't do it. You need more than XML. You need Web services. You need a public, well-understood, integration technology.

Dow Jones is also beginning to see many of its business partners and customers moving toward Web services.

A lot of our partners, such as Thomson Financial, have embraced Web services as a strategy for application integration and B2B integration. Additionally, big customers like Merrill Lynch, GoldmanSachs, and Morgan Stanley are all beginning to use Web services as a B2B strategy, as well as an enterprise application integration strategy. Some of those companies are, in essence, technology companies— much more so than we are. We're a content company. We have a lot of technology capability and expertise that enables our products. But, at the end of the day, the core mission of my company is business news for businesspeople, not technology.

However, when you move in to companies like Reuters, Thomson Financial, or Bloomberg, they can be considered technology companies. For example, at Thomson Financial, they have something like 2,500 employees in technology, because it's the nature of their business. Technology provides a competitive advantage, so companies like that are absolutely adopting Web services. I think that Web services have tremendous traction in these companies. So, if we want to integrate with CRM vendors, portals, or trading platforms, then we're going to need to adopt technologies that will make it much easier to do so. Otherwise, every customer is a one-off, and that's not good for either them or us.

Web services is a technique that gives you flexibility to integrate applications on the desktop, to integrate applications over a wide area network, and to integrate applications over a local area network, using open industry standards. And it also leverages your investment in all your existing technologies. Web services really are a paradigm shift in that they make use of open industry standards. As a result, the network is already there. They can truly offer you a integration path that's easy to understand, easy to implement, and easy to support. I think it's a winner.

Web services is the ideal technology for CIOs facing the dual challenges of simplification and cutting costs, says Godfrey.

Every CIO in America is facing the paradoxical challenge of cutting costs and raising productivity. One challenge is that we're all trying to simplify our infrastructure. We're all moving toward a portfolio orientation, and we're looking at our application inventory as a portfolio of assets. And we're all trying to figure out how to get more leverage out of that portfolio.

CIOs are all trying to simplify, consolidate, and rationalize. In the old days, you would maybe mount one of these big "boil the ocean projects." You know, doing CRM across the company, supply chain management, or enterprise application integration. Everything was enterprise-level with a big "E." They became very difficult projects. A better technique today is continuous improvement with the point of view in mind that says: Integrate and consolidate where you can prove there is incremental value to the company.

With Web services, you don't suffer the big productivity hit that the "boil the ocean projects" create. At the sametime, you can accomplish the goal of simplification. I think all CIOs are looking for a simplification dividend. The dividend comes from lowering your costs and reducing your operational head-aches. Because Web services are at a sufficiently high level of abstraction, you can leverage existing technologies; they sit on top of legacy. The power is in having a standard way to investigate.

When it comes to a content-driven company like Dow Jones, Web services allow easier content reuse, notes Godfrey.

I've got multiple systems to create The Wall Street Journal and the real-time Dow Jones newswires. As I share and repurpose content, everything is done through black boxes—-file handlers. The format coming out of The Wall Street Journal publishing systems is not compatible with Dow Jones newswires, so I have to write code. The format of the copy coming out of Barron's, which ultimately goes to WSJ .com, also isn't in the right format, so I have to write code.

Part of my competitive advantage at Dow Jones is tied to how well I can reuse content and tailor it to the medium in which it's ultimately expressed. Today, that takes programmers and big projects. With a Web services approach, I can make the repurposing and the reuse of content a simpler, and less expensive, process.

Godfrey notes that the mere mention of Web services tends to raise unnecessary apprehensions in some CIOs.

My belief is that Web services, as a programming technique, aren't particularly difficult. They represent nowhere near the paradigm shift that was presented by second-, third-, or fourth-generation languages. They're relatively straightforward. I believe you can get your feet wet with Webservices without a lot of risk. You can use them as an interapplication integration technique. That, I think, offers a lot of benefits right away. Using Web services as a means to build up a repository of reusable components, however, represents another level of maturity. That can be harder to tackle. Like object-oriented development in the 1980s, will the company stay the course long enough to actually create enough objects to make the investment worthwhile? It's not a strategy you want to go halfway with.

Still, Godfrey believes that Web services' potential for providing inexpensive, reusable components is strong.

The idea of reusable components and methods and object-oriented technology has been around for quite some time. Earlier attempts failed because they were swept away by new technologies, because the development skills were hard to come by, or simply because they weren't very good. One also had to wait a long time and build up enough components to make the effort worthwhile, and many people lost the will to wait that long.

For example, I was part of a Smalltalk team years ago when I worked in the insurance industry. We had built up hundreds and hundreds of objects, using Smalltalk for an insurance claim system. But the process just took too long—several years. Management lost patience with the fact that it took so long, and it was so darn expensive. And, at the end of the day, it was still proprietary. So we killed the project.

Web services are different, though. With Web services, you don't have to rewrite all the functional code. While some services will have to be written from scratch, many will be reusing capabilities you already have. That's how it's different from older models, which had to be fit to proprietary languages like Smalltalk, or proprietary frameworks, like CORBA.

Dow Jones's Web services platform marks a major change from the company's previous technology.

We didn't have a framework; we were, primarily, a C application shop with applications running on top of a variety of Sun boxes and an IBM mainframe. There were some COBOL applications on the mainframe and close to a dozen C++ apps running on top of the Sun boxes.

The use of middleware remains an unsolved problem for Godfrey.

We have NQ Series in some of our application portfolios, but we don't yet have a standard middleware platform across all applications. That's going to be a part of our next generation work.

We kind of look at Web services as a stand-alone topic, but it's actually integrated to the middleware in an N-tiered distributive model. We are continuing to build N-tier distributive applications. That's a given. A very small portion of our portfolio uses third-party middleware—in the past, we usually rolled our own. Going forward, we'll begin to use even more third-party middleware. That's being evaluated.

On the whole, Godfrey has found few drawbacks to Web services.

There were some implications where performance was considered an issue. There's some overhead with Webservices, with XML and SOAP. The XML vocabularies themselves, for example, are bulky. There's something like a two-times or three-times increase in characters transmitted. You need to be careful of that. So there was some concern, when we did our proof of concept, that Web services may not fit every usage profile.

Godfrey is particularly concerned about the suitability of Web services in a real-time environment.

Anyone who's engaged in real-time data management ought to figure out for himself if Web services make good sense. I think, ultimately, they will. But the nature of a real-time environment is that it's real time. People are paying a premium for a real-time application, and the quality, authenticity, and authority of the data are all important.

In real time, the publishing systems are typically built differently. The error recovery, tagging, coding, and indexing are all important. It's just a different paradigm. So, as we move to Web services for all the reasons we talked about, we should make sure that we don't have a service-level issue in real-time applications.

Another area of concern for Godfrey is XML.

There are a lot of people who can create XML, but it's often not well formed or well thought-out. You've got to be careful of that. I think it's very much like creating a schema for a relational database. You need people who are really good at information management and information models, which is not a technology problem per se. It's a matter of getting technology to relate at various levels of abstraction data to make sure that the XML implementation will actually work and also provide scalability. That's the issue. You can hard-code anything, but, during the day, you want to make sure that it's leverageable, extensible, and scalable. That only comes about if someone really smart thinks through the information model in such a way that it is scalable.

This goes back twenty years to how people would argue about normalization techniques. It's a little bit of the same thing now: How do you approach creating your information models? Is it a good fit for the problem that needs to be solved? That's something I don't know; I can't give you the definitive answer. But it's something I'd be a bit wary of—that you don't apply Web services to every problem. You still have to think through and ask what the right technique is for a particular problem.

Godfrey also has concerns about third-party UDDI services.

I'm certainly not going to be a market leader in that regard. I do think the UDDI concept makes good sense. I suspect for a period of time I'll provide my own UDDI directory. I wouldn't use a third party, but in the future I might, if it was a trusted third party, and if there were enough applets and components that it would really speed my development.

I'm not particularly worried about security, and I'm not particularly worried about the flexibilities of WSDL. I'm not particularly concerned with semantic applications—I think there's plenty of robustness in the protocols and the framework.

Godfrey believes that the excitement that Web services have generated is mostly justified.

Web services have provoked some new thinking for me. As a CIO, the debates and dialogues about Web services have resonated with me more than other enterprise application strategies. So, I'm more interested in Web services than I have been interested in some of the predecessor technologies and approaches. The idea of a services-oriented architecture is very appealing to me.

The Web services bottom line, says Godfrey, is the bottom line.

In the financial services space, Web services are getting a lot of traction. So Dow Jones has to keep up. It's our intention to use Web services to add value and to lower costs.