Making XML Work in Business
January 2, 2002
XML was developed to meet the real needs of real organizations, and its novelty and its promise have attracted plenty of attention from technical and business people. For XML to continue to thrive, however, it needs to deliver real value to companies and organizations, particularly in these tough economic times. Several of the sessions at IDEAlliance's recent XML 2001 conference showed how XML can deliver for businesses. But the discussions also suggested that the number of organizations able to take immediate advantage of XML is still quite small, and most businesses will probably not see benefits from XML until further down the road.
Information Management, The Early Payoff
According to several presentations at XML 2001, organizations can get the biggest returns from applying XML to their information management practices. However, the number of enterprises likely to benefit in the short term from this kind of investment in XML appears quite small.
Una Kearns of Documentum outlined the information management challenges faced by both public and private organizations. These organizations collect and generate massive amounts of information – what we now call content – such as catalogs, contracts, requests-for-proposals, product specifications, news items, marketing data, technical documentation, financial analyst reports. The content produced and collected in individual departments often ends up managed differently in various departments, which makes an overall organization-wide strategy difficult.
Despite the difficulty of the task, organizations able to tame this beast can reap significant rewards. The most immediate savings come from reusing information across divisional boundaries. For example, capturing a company’s product specifications in a way that can be reused directly by its marketing and service documents save the marketing and service people hours and days of time recreating that information and reduces the potential for inconsistencies and errors. Many companies have yet to take these first basic steps, so companies that produce, report, or manage large volumes of content have a good opportunity to reap significant savings early on.
The idea of information reuse is hardly new, especially to participants from earlier XML and SGML conferences, but the relative simplicity of XML should make it more palatable to larger numbers of companies and organizations. The ability with XML to identify key variables and identify them with common tags can make information in one department meaningful to other departments. As long as the people creating the content understand the idea – in itself no small accomplishment, as discussed in other XML 2001 sessions – a company has a fighting chance to save significant time and money managing its content.
Heavy Trucks, Heavy Benefits
Jonathan Parsons of XyEnterprise talked about how one of its clients, Freightliner, uses XML to manage content. Freightliner, a division of Daimler Chrysler, manufactures heavy commercial trucks and buses, as well as specialized heavy trucks such as fire engines and military vehicles. As one might expect, such vehicles are very expensive. Throughout the U.S., Canada, and Mexico, jurisdictions impose weight and height limits on trucks, as well as regulate the nature of the loads hauled by these vehicles. Thus Freightliner needs to customize its trucks for each customer.
This requirement means Freightliner has a complex information management challenge, as well as a demanding manufacturing task. The company configures each vehicle at the time of manufacture, which means Freightliner needs to capture data on each component in the vehicle and its properties from the marketing staff and transmit those specifications precisely to the manufacturing division. The information management challenge gets compounded by the need for customized owner manuals and service bulletins for each vehicle. With each unit of output customized for the buyer, there is no margin for error anywhere in this process.
Freightliner decomposes each vehicle into components and assigns a unique identifier – what it calls a "datacode element" – to each component. Each component then gets customized with attributes that reflect the product lines. These contextualized components (as they call them) are associated with pricing and weight data as needed. The components then get collected and assembled into documents – price quotations, manufacturing specifications, owners’ manuals, service bulletins – at the time of their publication.
For Freightliner, XML (and SGML before it) provides business value by making it possible for the company do business in this way. Without XML the company would likely find it financially impossible to design and deliver customized vehicles, along with the manuals and after-sales support. This case offers perhaps an extreme but valid case of how XML delivers value immediately to the company. Freightliner had used SGML previously for producing its service manuals, but the introduction of XML enabled the company to expand the idea of contextualized components to its marketing and manufacturing operations, thus decreasing the time needed for configuring and manufacturing the vehicle. Parsons said Freightliner plans to expand the concept to involve its supply chain partners.
Serving Financial Services
Financial services is another industry where information plays a vital part of the business. Bruce Sharpe of SoftQuad described the needs of their financial industry clients and how XML meets them. In the world of finance, content, for all intents and purposes, is the product. And timing is critical. Investment banks and brokerage houses were among the first companies to integrate the Web into business operations, and a number of them now use XML to create new information products and channels for delivery to the customers.
Sharpe said the Web did not automatically provide benefits to the investment analysts who prepare many of the time-critical advisories for the clients. Investment companies had long-established processes based on paper, but these processes generated errors and took valuable time. In many cases, the companies’ original web processes paralleled the paper-based operations, offering the analysts only more work, time, and errors and no real added efficiencies.
For financial analysts, the ability to separate content from presentation makes it possible to tag free text and data for future reference or reuse. In the often tight regulatory environment in which analysts work, XML helps provide better audit trails. When financial companies can organize their information into databases, XML can provide further benefits. The analysts can assemble the data into customized products for customers, using preformatted templates or stylesheets, giving customers more focused and personalized information products as well as faster service. For customers in the financial services industry, the institutions can syndicate the content and customers can republish the information on their own sites.
Sharpe listed the savings that investment houses realize as a result of XML. The main benefits are the increased efficiency of the highly-paid analysts. Companies can reuse data that would otherwise take valuable time from the analysts to research and recreate. The use of XML eliminates the manual processes held over from the hard copy days, as well as parallel processes introduced with the first web pages. And the customized information products create closer relationships with customers.
A Very Steep Pyramid
While manufacturers like Freightliner and financial institutions may realize significant benefits from XML, a closer look suggests that these companies have large, complex, and demanding information needs that make the investment in XML worthwhile. In both cases the use of XML has direct and significant financial implications, since each Freightliner truck carries a high price tag, and investment decisions usually involve significant capital outlays. Thus the potential number of companies or organizations able to realize these benefits is relatively small.
Counting these organizations is not easy, but assume for the moment that larger companies are more likely to need the information-management opportunities offered by XML. It is the larger companies which have operating divisions and greater financial and regulatory reporting responsibilities and are thus more likely in the market for XML information management systems. A look at the last U.S. economic census (1997) shows that the distribution of large to small companies resembles a very steep pyramid. American companies with 500 or more employees – the traditional cutoff between large and medium or small companies – number just over 16,000 or barely 0.3 percent of the total number of companies.
In fact, the vast majority of businesses in the United States are small. 89.5% of American businesses have 20 or fewer employees. (See "Statistics About Small Business and Large Business from the U.S. Census Bureau." http://www.census.gov/epcd/www/smallbus.html). XML needs to deliver value for the large number of smaller companies if it is to achieve sustained ubiquity.
E-business, for the Longer Term
Fortunately XML has another major application area that can offer this kind of business value. The hundreds of e-business frameworks, vocabularies, and protocols that have emerged over the past two years indicate an intense interest by many industries in taking advantage of XML’s ability to exchange structured data between companies and organizations. And some of the XML 2001 sessions provided a preview of the way businesses can benefit from these developments.
The conference devoted an entire day of its e-business track to financial applications. In this track, Rick Schumacher of Wall Street Systems discussed the Financial products Markup Language (FpML), which was designed for use with a type of investment called a "derivative". Derivatives are contracts based on expectations of the future value of often volatile financial measures such as interest rates and foreign exchange. No two derivative contracts are alike and each often needs individual negotiation.
Derivatives are traded privately, usually between institutions, not through an exchange. As a result, they do not have the usual regulatory protections found in trades made through the NYSE or NASDAQ. Information is the lifeblood of derivatives, so any news with a bearing on the contract is eagerly sought out. Again, timing is critical. The value of contracts can change markedly in hours or even minutes.
Schumacher said desktop technology made derivatives trading possible. With spreadsheets and databases, analysts at investment banks or brokerage houses could design and run financial models to help predict movement of interest or foreign exchange rates, to help reduce the overall risk of derivatives. But because the companies trading in derivatives had to keep these models secret, most of the interactions between companies were manual.
The industry thus faced the challenge of finding a way to describe a standard electronic derivatives document that would still enable participants to develop a complex and individualized contract for each trade. With the structured and extensible nature of XML, the industry could design a vocabulary to meet these conditions. The FpML organization is supported by 11 large investment banks and brokerage houses, as well as financial systems vendors and networks such as Reuters and S.W.I.F.T. Its current specifications cover overall architecture and interest rate derivatives, while it is working on foreign exchange and equity derivative products.
Do the Math
Tony Coates of Reuters described how the financial services industry benefits from FpML. The industry had already developed standard contracts for many of its products, what it calls vanilla contracts. With FpML, parties to trades can automate the parts of the contracts with which they had agreement, thus reserving human intervention for negotiating areas of disagreement. According to Coates, using FpML will help reduce the current three to five-day preparation time for derivatives down to one day by 2005, the target for the industry. And in an industry where timing plays such a critical role, this increase in turnaround time will payoff immediately for all participants.
One of the companies implementing FpML is SwapsWire, a network of 23 investment banks for the electronic trading and confirmation of financial derivatives. Guy Gurden of SwapsWire said the first product support by the company is interest rate swaps (thus its name), although it plans to support other kinds of derivatives in the future. Swaps are an agreement to exchange future cash flows. For example, one party makes payments based on fixed rates, while another party makes payments based on floating rates. These agreements can run for as long as 30 years, although three to five years is the normal period.
After the agreements are drawn up, they need to be confirmed. Today for confirmation, the trading partners exchange documents by fax. Either they can take one document and fax it back and forth, or each party draws up a document that gets exchanged and discussed. Even the simplest document can run several pages, so the process is time consuming and error prone.
Gurden laid out a few metrics to show the potential impact of FpML. Today, the average weekly derivative volume of a typical financial institution is 380 trades. Each support staff person can handle about 9 trades per week, which means the average institution needs some 42 staff people to process the trades. Each support staff person has an average loaded salary of $87,000. FpML can increase the throughput for each of the staff people and reduce the negotiation load on each of the analysts. As they say on Wall Street (and elsewhere), do the math.
But SwapsWire will likely have its greatest impact on the improvement of the business processes used in swaps. Today, confirmation takes place after the parties agree to contractual terms. The traders match the provisions of each document, and once they are all in agreement, then the contract is confirmed. By bringing the traders on the SwapsWire network, the parties can agree to the terms of the contract during the trading process, thus cutting out much of the time needed for confirmation.
Another session in the track featured the Extensible Business Reporting Language (XBRL). Zach Coffin of KPMG described how XBRL could well become the catalyst for an entirely new way of reporting business performance. XBRL started as an accounting vocabulary, but now encompasses most of the ways a company reports its performance or activities to investors, government agencies, or the public.
As an example of the need for XBRL, Coffin described the potential benefit for one of KPMG’s banking clients. That bank, according to Coffin, is one of the best in the world at processing loan applications, but it still takes nearly two days on average to turn around each application. An analysis of the bank’s processes showed its staff spent 90 percent of this time on data discovery and paper-shuffling mechanics, and only 10 percent on real decision making. Better management of the information from credit customers, much of it tied to business performance, could reduce time spent on loan applications and increase the throughput of loans significantly.
Coffin said XBRL designed its vocabulary using the traditional supply chain as a model, with each piece of information like a bar-coded and uniquely identified inventory item. Using this model, XBRL aims to better manage the flow of information through the business reporting supply chain, from internal business operations, to management reporting, to external public reporting, and finally to investment and lending analysis. Examples of current uses for XBRL include accounting, financial reports, loans, mutual funds, and economic indicators. The XBRL organization is working on regulatory filings, tax filings, contractor ratings, and performance-related press releases.
Extending the EDI Experience
Joe Smolic, an independent consultant, described how the book publishing industry plans to use XML to extend its previous work with electronic data interchange (EDI) beyond the few large companies to the much larger universe of smaller companies who cannot afford the high cost of EDI.
Smolic said book publishers found that to buy the paper and services needed to manufacture books, hard copy purchase orders cost from $65 to $85 each, while electronic purchase orders using EDI cost $1 to $4 each. He added that electronic transactions will reduce errors in the complex transactions with manufacturers and paper suppliers, as well as reduce the time spent on paperwork.
Less tangible, but still meaningful benefits include better management of the production process from more accurate and timely information on the status of production jobs, paper usage, ship notices, and invoices tied to the individual print jobs. Smolic added that the use of XML can also encourage new business opportunities, such as opening up new channels to the end-consumer, including direct sales from the publisher, on-demand book printing (no more out-of-print books), personalized book production, and selling books incrementally, chapter by chapter.
While these benefits to businesses, including small ones, can be significant, for the most part many of the systems discussed at XML 2001 are still being planned or projected. Several conference sessions and tutorials addressed Electronic Business XML (ebXML) and web services as infrastructures that can deliver these benefits to smaller businesses. One of ebXML’s main goals is to make e-business possible for smaller businesses, and both Coates and Smolic mentioned ebXML as a framework for their respective industry vocabularies.
The challenge for XML is to make e-business real for smaller businesses. Not only can smaller companies and organizations benefit, but XML vendors can start showing the value that they have promised since the beginning.