Reader ROI\n\nSee how an entrepreneur marshaled forces of IT to realize his business vision\n\nLearn what it takes to link disparate systems in a B2B marketplace \n\nUnderstand how managing staff can be vital to an exchange project\u2019s successMichael S. Levin revels in what some might call the extreme. He doesn\u2019t just hike, he climbs mountains, including the Chugach in Alaska and ranges in Wyoming\u2019s Jackson Hole. He doesn\u2019t just sail, he\u2019s a past skipper in the Admiral\u2019s Cup and Southern Ocean Racing Circuit as well as the recipi- ent of the Stamford Yachting Club\u2019s De Coursey Faeles Trophy. Forget recreational skiing\u2014Levin, who is 50, favors couloirs, the steep cliffs that are no stranger to avalanches. So what\u2019s next for the thrill-seeking Levin? Applying his love of adventure to conquering the rough and tumble world of business-to-business e-commerce.As chairman and CEO of e-Steel, an online marketplace for the steel industry, Levin\u2019s certainly picked the right place to test his endurance in the face of a challenge. On one hand, the 2-year-old e-Steel (www.e-steel.com) is considered a veteran and one of the more stable players in B2B, which is expected to be the next frontier for blockbuster technology growth in e-business. Its own revenue model calls for e-Steel to receive a fee\u2014paid by the seller\u2014each time a transaction is complete. The company also hopes to boost revenue by building and selling specialized applications for its exchange for specific industries (like automotive) and by selling branded e-Steel applications for other marketplaces.The fact that e-Steel has a multipronged strategy is key. With investors and pundits still reeling from last spring\u2019s market crash and shakeout among business-to-consumer Internet sites, B2B marketplaces are under the gun to provide value-added services that go far beyond streamlining procurement if they have any hope of long-term survival. The focus now is on integration, or the process of linking an exchange to participating companies\u2019 back-end financial, order entry, inventory and manufacturing systems. The goal: to create a highly automated, online supply chain that delivers such efficiencies as reduced transaction costs, less inventory in the pipeline and improved collaboration, forecasting and scheduling among suppliers, suppliers\u2019 suppliers and so on. It\u2019s a tough game, and while Levin claims New York City-based e-Steel is certainly ahead of the curve, he admits it is far from a master of integration. "It\u2019s been our total focus in the last few months," says Levin. "Integration is paramount in thinking through all the new developments in the company." This relentless focus on integration is starting to pay off. In July, e-Steel inked a three-year, multimillion-dollar partnership with Australian-based Broken Hill Proprietary (BHP) Co.\u2019s steel business\u2014the 19th largest steel producer worldwide\u2014to build a custom network powered by e-Steel technology. Initially, the network will be open to BHP\u2019s Australian customers and later expand to key markets internationally.Integration at the CoreStrategically focusing on integration is a good move for e-Steel. By many accounts, revenues from B2B e-commerce are poised to balloon to the hundreds of billions, maybe even trillions of dollars over the next few years. But experts say those rosy forecasts are predicated on making integration a core capability of B2B exchanges. "Within two years, B2B e-commerce will hit $900 billion, but that\u2019s peanuts compared with the size of the economy," notes Tom Harwick, research director for supply chain management at Giga Information Group in Cambridge, Mass. "It\u2019s going to grow to a much bigger figure by 2005, based on the assumption that integration happens. If it does, B2B will be the preferred way of doing business. If it doesn\u2019t, growth will almost certainly stall." All the hoopla surrounding B2B\u2019s potential has fueled the perception that a lot of the integration work has been done. Not so, say the experts. According to a June 2000 report from Forrester Research, only four out of the 50 large companies surveyed that are involved in e-business have integrated their back-office applications with electronic marketplaces, although most (85 percent) have integration plans in development. Then there\u2019s the notion that building bridges between an exchange and customers\u2019 back-end systems will be relatively straightforward. Far from it, when you consider Forrester\u2019s findings that 70 percent of companies plan to participate in more than one marketplace, most targeting between four and five. "Unless you talk to the people who are really in it, there\u2019s a perception that more integration has been achieved than what has happened to date," says Giga\u2019s Harwick. Integration issues have taken a backseat because most of the freshly minted exchanges believe they first need to focus on building up membership and liquidity in terms of the number of transactions. And that process has just begun. Forrester\u2019s research shows that only 36 percent of all e-marketplaces have done more than 100 transactions a month. "The real issue is that there\u2019s not enough traffic through the sites yet, so the motivation isn\u2019t yet there [to focus on integration] on the part of the companies that build marketplaces and the companies that join them," says Simon Yates, a Forrester analyst and author of the report, "B2B Integration Road Map." Besides the technical issues associated with syncing up computer systems, B2B integration, if done properly, hinges on rethinking global supply chain practices and instituting massive change not far from the scale of upheaval associated with the 1980s reengineering fad. Getting companies to cast a critical eye at their existing processes and interaction with their supply chain\u2014not to mention, to be willing to make modifications\u2014is where most of the suffering will lie. "To get us to nirvana, the early adopters and mainstream adopters are doing the work that\u2019s required over the next couple of years," Harwick says. "They\u2019re bearing the pain but will reap the benefits of being early movers."First Come, First ServedE-Steel is clearly positioning itself as one of those integration pioneers. Yet while Levin says the idea of linking the exchange to a steel company\u2019s back-end systems was always part of the e-Steel big picture, it was not the focal point when the site was conceived in early 1998 or at the time of its inaugural transaction on Sept. 7, 1999. That\u2019s when Worthington Steel Co. purchased several truckloads of prime hot-rolled coils from Cargill Ferrous International.At the start, the thinking was to create a networked, global marketplace that would alleviate some of the inefficiencies in the steel industry, which was struggling with low margins and cumbersome, paper-based processes. It had been Levin\u2019s goal for years, having started in steel at his stepfather\u2019s company, Titan Industrial Corp., after graduating from Harvard Business School in 1974. Although no one expected him to stay in the business for any length of time, Levin saw promise in what many viewed as a stodgy, unprofitable industry. He became addicted to the global marketplace that sent him to exotic locales like Turkey and India, so much so that he eventually ended up buying Titan from his stepfather and logging over 25 years in the business. "Steel mills are the fundamental building blocks of our economy," he explains. "Steel is on a scale where everything is big. It\u2019s everyone\u2019s notion of what industrial power is all about."Levin\u2019s obsession first resulted, during the early 1990s, in something he called Steelnet, a global network where participants could communicate and transact business using satellite communications. While the idea for Steelnet stuck, the company never materialized since satellites were not robust enough at the time to deliver the real-time information required for Levin\u2019s vision. When the Internet took off years later, Levin revisited the concept. This time, he put up about $1 million of his own money in initial seed capital and brought on financial and marketing gurus to help make refinements to what he was now calling e-Steel. They got funding in 1999 from the cream of the venture capital community, including Kleiner Perkins Caufield & Byers, Bessemer Venture Partners and Greylock, and the race to B2B e-commerce was on. "This was not greed-driven," Levin says. "In my mind, this was a culmination of a career in steel and a way to make a difference in the industry."Like most early exchanges, Levin and crew initially believed e-Steel\u2019s contribution would be to bring buyers and sellers together more efficiently over the Internet and to open up doors to trading partners that were otherwise out of reach. Forget the laborious and error-prone fax and phone process traditionally associated with buying and selling steel: An online exchange model lets industry players from steel mills to service centers submit requests for proposals for all types of steel products, compare pricing and packages across multiple suppliers, negotiate price and complete transactions, all from a secure, global marketplace. Unlike most of its competitors, including primary rival MetalSite (www.metalsite.com) of Pittsburgh, e-Steel took aim at negotiated transactions among known parties instead of relying on auctions for spot purchases of goods. Steel makers LTV Steel Co., Steel Dynamics and Weirton Steel Corp. were the initial investors in MetalSite when it launched in 1998 as a limited partnership.As e-Steel began rolling out to beta customers in the summer of 1999, the thinking surrounding integration rapidly changed. Instead of appealing to companies for one-time transactions with new partners, major steel companies were looking at the e-marketplace as a venue to transact all of their business with their largest suppliers and customers. Given the scale of possible transactions, e-Steel would not be able to deliver any significant value beyond the old way of doing business via fax and phone without a solid integration strategy. "The concept of integration was always in the back of our minds because once we built a marketplace, we knew people would want to eventually connect it to ERP [enterprise resource planning] systems," Levin explains. "But we had to put it front and center because to the extent you don\u2019t connect, you\u2019re left with little more than e-mail as a replacement for fax and phone."To kick start the effort, Levin brought in Chief Technology Officer Tom Costello, a former senior vice president at CyberCash, a maker of e-commerce payment software. Costello joined last October, just a month after the first e-Steel transaction. At the time, there were only five e-Steel employees and a couple of dozen folks from Computer Sciences Corp. (CSC), a consulting company that built the initial site using packaged e-commerce software from BroadVision. Costello soon recognized that e-Steel had considerable work to do to reposition itself around integration. "When I came on board, we had nothing from an integration standpoint," Costello recalls. "We had to take control of the technology that would drive forward our vision."Quickly, Costello went about piecing together what he refers to as a "building-block" integration strategy. First, he built up an internal IT organization so that e-Steel could work on some of its own integration technology, along with CSC. He also determined that the BroadVision e-commerce platform was not scalable enough to support a dynamic trading exchange with hooks into back-end systems. As a result, e-Steel redesigned its architecture and chose BEA Systems\u2019 WebLogic transaction framework to replace the BroadVision software. It also sought a partnership with webMethods of Fairfax, Va., which makes hub and spoke software that delivers a secure, dynamic link between the e-Steel exchange and a company\u2019s back-end systems\u2014using a wide variety of standards, such as extensible markup language (XML), common delineated files and electronic data interchange (EDI). In addition to the relationship with webMethods, Costello helped ink a partnership with USX Engineers and Consultants (UEC), the Pittsburgh-based systems integration arm of the U.S. Steel Group (part of USX Corp.), to provide integration services to steel companies looking to participate in e-Steel. Through its ValueTrack service offering, e-Steel also works with companies to assess their e-business goals, and it develops and deploys an integration plan. There are two other pieces to Costello\u2019s building-block integration architecture. To help automate the process of loading inventory information into the marketplace, e-Steel built DataJet\u2014a free data mapping and uploading tool. DataJet helps customers upload made-to-order, current inventory and product catalog information to the site without data format conversions. E-Steel is also developing and rallying industry support for the Steel Markup Language, a set of extensions to XML that would provide a common format for sharing information among players in the steel industry.MetalSite, e-Steel\u2019s big rival, has also recognized the need to address integration, of course. In December 1999, MetalSite completed a first-phase test of its integration capabilities with Bethlehem Steel Corp.; this allowed EDI data on the steel maker\u2019s products to be automatically transferred to MetalSite\u2019s Web catalog without rekeying the information. While this capability is similar to e-Steel\u2019s DataJet tool, e-Steel\u2019s adding the webMethods technology takes its integration capabilities to the next level, experts say. DataJet automates the transfer of product data from a seller\u2019s system to e-Steel\u2019s catalog, while the webMethods middleware, once installed, maps both data and business processes so that information gets passed among the disparate systems of parties involved in the exchange, automatically, in real-time.Together, these pieces give e-Steel a solid foundation for integration. "Since day one, e-Steel has had a more robust concept of what\u2019s required," notes Thomas Abrams, an equity analyst who covers the steel industry with Credit Suisse First Boston in New York City. "But to some degree, the battle has just begun." Integration with everything that\u2019s required for a transaction\u2014from sharing inventory data among suppliers to work-in-progress reports and production data\u2014won\u2019t happen at e-Steel or any other e-marketplace for another 12 to 18 months, Abrams says.Laying the GroundworkCostello would have to agree. Even with all the early groundwork, he admits e-Steel is just getting into heavy lifting as far as integration is concerned. Most of the 3,500 e-Steel members are now just beginning to use the DataJet mapping tool as a way to get inventory from their back-end systems posted to the site without manually rekeying. And as a one-way pipe into the marketplace, DataJet is really only an interim solution, offering a level of integration far less than what large steel players are after. E-Steel\u2019s real differentiator, Costello says, will come with the webMethods software, and that effort\u2014which currently requires months of integration work with consulting partners\u2014is only beginning at a handful of sites.U.S. Steel in Pittsburgh, which is a minority stakeholder in e-Steel as well as a beta customer, is furthest along in integration, having used DataJet and started work with UEC as an integration partner for the webMethods software. Since February, when it made the decision to partner with e-Steel, U.S. Steel has done limited transactions on the site, mostly with nonprime sheet products\u2014mill-downgraded materials, secondary products or excess prime offerings. U.S. Steel\u2019s use of e-Steel has been limited mostly because it requires a high degree of integration to do any type of volume transaction, according to Robert McClintock, commercial systems manager for U.S. Steel.Specifically, U.S. Steel needs to link proprietary manufacturing resource planning systems in each of its four plants, which run on IBM mainframes in Pittsburgh, to e-Steel so that materials available for sale could get automatically posted to the exchange. "To do any volume, we need to get inventory seamlessly transferred to e-Steel\u2014the overhead associated with manually inputting data into the site would make it not worth our while," explains Gene Trudell, U.S. Steel\u2019s general manager of computer services. Integration on the back end is also required once a negotiation is consummated. There, U.S. Steel needs to be able to take the result of the deal struck on e-Steel and flow it through its back-end order entry systems built on Oracle Corp.\u2019s Oracle Applications so that it can be invoiced, billed, shipped and so on, like any other transaction. Finally, to maximize efficiencies, there needs to be integration at the customer level so that U.S. Steel\u2019s partners can seamlessly upload their purchasing requirements to the exchange.McClintock and Trudell are confident that by working with e-Steel and UEC, they\u2019ll be able to achieve the right levels of integration. However, they know it will take time and require significant development resources on their part as well as a willingness to reevaluate business processes. For instance, while DataJet addresses volume uploading of data to e-Steel, U.S. Steel realized it had no easy way to present inventory data for transfer since it resided in four separate plant systems. As a result, U.S. Steel decided to change things on its end. With the help of UEC, it\u2019s designing a common inventory database, from which U.S. Steel can apply business rules and let the webMethods software automatically handle the data collection and transfer to e-Steel. "You have to understand how you do business and take your understanding of that and objectively decide whether it\u2019s the right way or if you need to change," McClintock says. Some of the integration pieces were put in place last summer to do volume transactions on nonprime sheet materials, and he expects to have pilot projects underway with select prime-sheet customers by the end of 2000.At the $3 billion National Steel Corp., another early e-Steel customer, full-scale integration is going to take the better part of a year. The team there is working with e-Steel and integration partner UEC on determining how to adapt its supply chain processes to make the most efficient use of the exchange, according to John Davis, National\u2019s general manager for information services and engineering in Mishawaka, Ind. "We\u2019re truly reformatting our business model to take advantage of this new model, and no one can do that in two or three months," he says. "There is too much cultural change\u2014all the things no one likes to talk about." (See "Culture Club," above.)As of August, National had conducted numerous pilot transactions with customers over e-Steel. As pilots with additional customers come online over the next few months, it will continue to refine its integration needs. "We\u2019re trying to figure out the touch points, where to hand off transactions and where it makes sense to integrate," adds Chuck Erickson, a systems engineer. Erickson says National has started to experiment with DataJet, but like U.S. Steel, he believes the real value of integration will come with the webMethods software. At this point, National is working with UEC to create new business processes to best leverage that software.It was e-Steel\u2019s ability to take the lead on development, including integration, via its ValueTrack program that convinced Ford Motor Co. to strike a partnership with the exchange to Web-enable its steel supply program. The alliance, which covers the auto giant\u2019s global Tier 1 suppliers, including stampers and steel sources, is aimed at eliminating manual processes and giving the complete supply chain access to the same database of inventory, order and pricing information. "This will allow us full integration," notes Karen Kish, purchasing specialist in Ford\u2019s raw material supply program in Dearborn, Mich. Ford and e-Steel are in the prototype phase. Ford has essentially committed to providing its business processes to e-Steel and has devoted a couple of information technology professionals and Kish, as program manager, to work on the project; however, e-Steel and its integration partners are doing the bulk of the deployment and project management work. The effort, which is slated to go live in the first quarter of 2001, is separate from Covisint, a forthcoming online exchange for the automotive industry announced as a Ford-General-Motors-DaimlerChrysler partnership.Given the scope of work required for integration, most big e-Steel customers\u2014including National, Ford and U.S. Steel\u2014don\u2019t expect to take part in additional steel exchanges unless they become a key requirement for customers. But while many place their bets on e-Steel, some experts caution that a one-horse integration strategy can be risky. "You have to play the game that the strategy could change tomorrow if the market changes\u2014you have to be flexible so you can get in and out of marketplaces at will," warns Forrester\u2019s Yates. "It\u2019s all about risk management and not putting all your eggs in one basket."What\u2019s risky business for some, however, is passion for e-Steel\u2019s Levin. And with companies like Ford, National and U.S. Steel in his camp, he\u2019s ready to take on the challenge. For now, couloirs and the sailing circuit might have to wait while Levin gives B2B e-commerce his best shot.