U.S. Steel Tries Tech Alchemy

By Mel Duvall  |  Posted 2002-06-17
U.S. Steel Base Case
Headquarters: 600 Grant St., Pittsburgh, PA. 15219
Phone: (412) 433-1121
Business: Integrated steel manufacturer
Financials: $6.38 billion total revenue; $218 million net loss (fiscal 2001)
CIO: Gene Trudell, managing director of information technology services
Challenges: Lower costs to better compete with foreign competitors, increase share of high-end steel market, return to profitability, maximize shareholder value with new business initiatives, lead consolidation of U.S. integrated steelmakers
Baseline Goals:
• Increase revenue per pound on steel products, from $383 a ton
• Reduce cost of manufacturing steel, by $10 per ton per year
• Reduce materials and parts on hand, from 20 days worth of inventory
• Turn Straightline distribution arm, which lost $7 million in the first quarter, profitable

U.S. Steel thought it knew what it was doing. Then, Ford Motor Company said USS ranked dead last in performance among its chief suppliers. But now the biggest American steelmaker would have the world—and the Bush administration, which is trying to save the company with fresh tariffs on products of foreign rivals—believe it is one of the most efficient producers around. Is it?

Buying software isn't always better than building it. Compare the options using our three-step tool.
  • Excel file (Compressed)
  • PDF file
  • In the summer of 1996, U.S. Steel got a wake-up call. Ford Motor Co., one of its three biggest commercial customers, was threatening to end its 70-year relationship with the largest steel maker in North America over a series of foul-ups involving ready-to-stamp steel.

    Massive coils of finished steel, ready to be turned into the shells of automobiles at Ford's facilities, were showing up without notice.

    The fiasco reinforced U.S. Steel's ranking as dead last among Ford's suppliers in delivering on promises. "We were put on warning. Unless we did something fast, we were in danger of losing Ford's business,'' says Chief Information Officer Gene Trudell. "It was that serious."

    The crisis forced U.S. Steel to adopt a bunker mentality. Where six years ago, individual businesses and factories controlled their own information infrastructures, the biggest steelmaker now has deposited the vast majority of its computing power in a fenced, low-rise Pittsburgh compound. The company also has led the steel industry in deploying new digital systems for interacting with suppliers and fulfilling the orders of buyers such as Ford, General Motors and General Electric. Now, U.S. Steel can track an individual slab of steel as it comes out of a casting facility in Gary, Ind., and follow it all the way to the customer's door. Not only does the steelmaker deliver the finished product when the carmaker wants it, it can confirm it made the delivery as ordered.

    U.S. Steel's transformation from technology laggard to technology leader was so thorough and rapid that Ford named it supplier of the year just two years after the wake-up call. But the $6.4 billion-a-year maker of the steel that undergirds the North American construction, automaking and appliance industries is hardly out of the woods. Its chairman, Thomas Usher, has been at the forefront of calling for protection from foreign imports. On average, the company last year lagged behind its toughest competition, Korean steelmaker POSCO, by an average of $68 a ton in overall costs.

    Its biggest disadvantage? Labor costs. That accounted for $66 of the difference. Part is due to the higher cost of union labor. But the biggest portion—upwards of $40 a ton—comes from health care benefits paid to retirees. That's $40 a ton POSCO doesn't worry about.

    As a result, even with the 30% tariffs on imports of flat-rolled steel products imposed by the Bush administration in March, U.S. Steel's costs still exceed those of its prime foreign competitors, including POSCO and Arcelor of Europe.

    Daniel Ikenson, a senior trade policy analyst with the Cato Institute, a Washington, D.C., think tank, says U.S. Steel is certainly one of the more efficient American steel companies. But it now lacks the size or scale to compete with foreign counterparts. U.S. Steel has the capacity to produce about 14 million tons of steel per year, ranking it about 10th in the world. That's less than one-third of the production capacity of Arcelor, the world's largest producer, and less than half the capacity of Korea's POSCO or Nippon Steel of Japan.

    "U.S. Steel's biggest problem is it doesn't have the same economies of scale that its foreign competitors have," says Ikenson. His solution: consolidation of steelmakers into larger, more efficient players. Not bailouts.

    But even at its current size, U.S. Steel is not inefficient. Indeed, World Steel Dynamics, a statistical service, estimates U.S. Steel takes just three hours of a person's time to produce a ton of steel. That's well below the 4.8 hours used in South Korea—and the 34 in India.

    Through the mid- and late-1990s, U.S. Steel in fact improved its systems to the point where it thinks it can run a profitable business peddling logistics of steel manufacturing and distribution to its domestic competitors.

    That new-found expertise last year led to the creation of Straightline, a new division that sells steel products directly to smaller customers, much like Dell Computer bypasses distributors to sell directly to individuals and small businesses. Another sales and services subsidiary, USX Engineers and Consultants, markets order fulfillment and supply chain management software not just to other steel companies, such as Inland Steel and National Steel, but aluminum makers and companies in other similar fields (See "Profiting from Information Technology" at www.baselinemag.com/uec).

    Developing technology is one of U.S. Steel's key methods of beating back domestic innovators like Nucor, that use scrap steel as low-cost raw material for its mini-mills. That forces U.S Steel to transform coal and iron ore into higher-grade products for automakers, skyscraper-builders and the like. Only if it can deliver premium products efficiently will it earn a profit against competitors with lower raw material and labor costs. In effect, U.S. Steel's main weapon at this juncture in its struggle for survival is no longer integration, but information.

    "Our competitive edge is not the equipment or the facilities," Trudell says, "it's the data we wrap around everything."

    Stacking the Decks

    The problem: It may be a fleeting advantage. According to Richard Fruehan, a professor of metallurgy at Carnegie Mellon University, it's next to impossible to establish a strategic advantage in the steel business with either information technology or sophisticated mill equipment. "Everybody has access to the same technology," he says. A company only needs to spend capital to bring systems in line with competitors.

    Trudell and his team, though, have tried to go a step further. Into the machinery in its factories and computers on the desks of office workers, it has threaded sophisticated systems for order processing and fulfillment, inventory tracking and logistics, and customer services. These systems have boosted plant productivity, increased office efficiencies and enabled U.S. Steel to launch Straightline.

    For instance, many of Ford's problems with U.S. Steel turned out to be faulty application of information technology. Systems were not in place to keep track of steel as it changed from raw materials to finished products; and they certainly were not in place to communicate just where along the manufacturing and delivery process an order for a particular customer happened to be. The quality of finished products even turned out to be dependent on information: customers would specify exactly what products they wanted, but specifications of their orders didn't get to the factory floor intact.

    "In making steel, you're always changing your grade and finding faster and better ways to make steel," says Jeff Davies, U.S. Steel's director of plant information technology services. "But, for the most part, I think it was a matter of communicating more directly and completely with the automotive customers.''

    Take the matter of advanced shipping notices (ASNs), a memo that lets a customer know what steel is going to be delivered at what time. In order to provide notice to Ford, U.S. Steel needed to collect status information in a timely manner from its processors—independent companies that work steel coils into finished products.

    U.S. Steel employs more than 120 outside processors, about 35 to 40 of which work on Ford products. Each processor used its own tracking systems, its own inventory codes, and its own tools for generating electronic ordering forms. Such electronic forms would be sent to U.S. Steel over a dialup system, known as an Electronic Document Interchange network. To take the data from a single processor, convert it to usable form and send it on to Ford would take U.S. Steel about 90 minutes—for a single message.

    That made it hard to satisfy Ford. The carmaker wanted ASNs to arrive before truckloads of steel showed up at its door.

    U.S. Steel was sending advance notices, in batches, on an hourly basis. However, many of these processing plants were strategically located near Ford sites—sometimes just a 20-minute drive away—which meant that even if there wasn't a problem with data collection, aggregation, or transmission, the truck could still show up before the ASN.

    That made Ford inefficient. If a truckload of steel shows up without an ASN, Ford employees have to manually collect details about the delivery and enter that data into Ford's systems. That canceled the benefit of technology. Inputting data is labor intensive and introduces human error.

    U.S. Steel knew it needed to make a lot of improvements, and quickly. "We put on a full-court press," Trudell says.

    The company wrote a new program to track inventory as it moved through the outside processors. The new system is "event driven," meaning that the receipt of an incoming message from a processor automatically triggers the next step in the chain—an update of the inventory system, the formatting of a shipping notice, or the sending of an ASN.

    U.S. Steel also incorporated electronic document formats standardized by the American Iron and Steel Institute, a steel industry trade group, and brought in a Sterling Commerce tool called Gentran to help translate messages.

    The upgrade was done, but U.S. Steel wasn't finished. U.S. Steel started to compile statistics on data errors sent by its outside processors. And, for those processors that failed to show improvements, U.S. Steel would set up training sessions with the processors' staff.

    Now, messages take 12 minutes of handling each, more than an 85% improvement.

    That gave U.S. Steel a better sense of what was happening with supplies—and a solid network for communicating with processors. That led to Straightline.

    "The process we developed to track these messages became the heart and soul of Straightline," Trudell says.

    Straightline is an attempt by U.S. Steel to earn a profit selling steel to smaller buyers. Large steel buyers, such as Ford, purchase directly from a big manufacturer such as U.S. Steel. But lower-volume customers—like a file cabinet maker or air conditioner manufacturer—look to distributors to take their orders and deliver products.

    With Straightline, buyers place orders over the Web or phone. Straightline manages customer inventories, provides up-to-the minute order status, and handles shipping from the mill, to the processor, and on to the customers.

    A Framework of Steel

    U.S. Steel's information technology budget is 1.1% of the company's annual revenue—about $70 million. That covers everything from the servers at headquarters, to PCs on workers' desks, to automation in the mills. By comparison, the median share of revenue spent on information technology by research and engineering companies is 9.2%, according to researcher International Data Corp.

    "We're frugal, or conservative, in our investment," Trudell says.

    The $70 million rarely gets spent on amenities. The data center team is hunkered down in a warehouse-style building, just a few miles south of headquarters, with the general luxury of an army barracks. Off-white walls blend into each other. Meetings are held in a windowless room with plastic chairs. Inside the computer room, older equipment—such as an early IBM 390 mainframe and a Storage Technology 4400 memory device—dominates the ends of the raised floor.

    "We can't afford to be revolutionary," Trudell says, "but we can be evolutionary."

    But even that has meant a lot more than simply making sure automakers get advanced notice of shipments. In the late 1990s, U.S. Steel undertook several ambitious projects—in order fulfillment, order status, inventory tracking—that eventually led to a homegrown method of managing suppliers. That, in turn, laid the foundation for Straightline.

    Many of these projects, designed to improve customer service, were done internally and with small expenditures. Decades-old facilities and equipment would get upgraded, not replaced. Existing technology would be built upon, not discarded. For instance, when the company wanted to give employees online access to pension information, it created access through a Web browser to its 20-year-old pension tracking system, instead of buying new software.

    A New Order

    Tom Zielinsky, senior director of information technology strategy at competitor Weirton Steel, for instance, figures that how a steel company takes care of its orders is one of the clearest ways it can separate itself from its competitors. Zielinsky says that when a steel company gets an order right—when it smoothly handles order entry, produces the exact quality of steel the customer wants, and delivers the product in the quantity requested and at the right time—that company is more likely to win repeat business.

    "You need a way to differentiate your business," Zielinsky says, "and I think you can do that with repeatability of customer orders."

    At U.S. Steel, David Sherwin, the company's director of order fulfillment, knew that the prime objective should be to allow customers to place orders electronically, specifying the products and quantities they wanted, with the system automatically figuring prices and shipment dates.

    "Everyone was producing the same steel," says Sherwin. How to fulfill orders would be different.

    But taking orders electronically is complicated, even in a seemingly stolid industry as steel. In addition to quantity, price, and delivery date, information on each steel product's composition, size and thickness must be calculated before an order is produced and shipped.

    "The whole methodology of coming up with an order that was correct from a metallurgical standpoint was a very difficult process," Trudell says.

    U.S. Steel tried to get started with an order fulfillment part of an Oracle software package. But then it added a product configuration system from a company called Concentra (which Oracle bought in 1998) that it thought could handle the intricacies and peculiarities of ordering a product like steel. Such orders require blending of raw materials and the manipulation of heat, tensile strength and other physical factors.

    The system could crunch metallurgical rules, production capability and limitation information, and the customer's exact product specifications. There was just one problem: The system needed to be populated with U.S. Steel product and metallurgical information.

    This turned into a lot more than U.S. Steel anticipated. "We underestimated how time-consuming and tedious the task of extracting and understanding the business rules and procedures would be," Trudell says. Many order processing business rules and procedures were encapsulated in thousands of lines of code on the company's mainframe computers. These had to be built into the configuration software.

    Software engineers needed to sit down with engineers and develop programs that could deal with the intricate mix of product specs and prices needed to serve customers. "This required thousands of hours of interviews and logic revisions before the application was considered acceptable," Trudell says. "Much of the knowledge was resident in the minds of our metallurgical engineers."

    But, when done, the product configurator put U.S. Steel at the forefront of handling customers' orders with intelligence—and fewer hands. Order changes—to compensate for inaccurate order entry or to redo an order taken that a mill couldn't handle—have dropped by two-thirds.

    Status Seeking

    A piece of steel that leaves the mill may go through as many as five different companies that treat, trim, shape and otherwise create a finished product born of steel. Customers, obviously, want to know where in this chain of processors their products are.

    Using Oracle databases and its electronic document network, U.S. Steel began to build a system that could pull together the locations of individual products and entire orders, from its plants and those of its processors.

    With ordering more accurate and the status of orders visible to customers, U.S. Steel had the beginnings of a system for managing its supply chain that could deliver tangible benefits to Ford or other big customers.

    Its information systems could "handle the capture of orders, the pricing of orders, the interaction with the plant—and we were able to actually create a promise date for the customer as the order was being entered," Sherwin says

    In effect, there could be no greater advance notice than telling a customer when an order would arrive at the moment an order was placed.

    This was radical. In the 1990s, most manufacturers, including steelmakers, were just beginning to try to implement supply chain software to smooth out distribution. Many companies, according to the Boston Consulting Group's Michael Shanahan, bought quick-fix, third-party packages that were force-fitted into their existing systems.

    Off-the-shelf software wouldn't satisfy U.S. Steel, though, says Shanahan. So Big Steel became a software developer.

    It tied its order fulfillment system into i2 Technologies' Factory Planner, a forecasting tool. Next came the homegrown order-status system. Then U.S. Steel mixed in two existing homegrown, mainframe-based systems: an inventory control system called iTrac it developed to keep tabs on material as it moved through its production facilities; and an automatic, order-generation system called the Mechanical Item Generation System, or MIGS.

    MIGS was developed in the early 1990s to reduce coil inventory at the company's mill in Gary, Ind. If Gary could better-forecast demand for finished goods at a customer's location, such as fenders and doors, then it could reduce the inventory it needed to keep on hand.

    MIGS, an IMS database application written internally in Cobol, was designed as an automated order entry sys- tem for repeat customers. By better understanding their demands, the company found it didn't need to keep as much steel on hand. In early usage, MIGS reduced inventories from approximately 33,500 tons to 24,000 tons. The system, which has been tweaked over the years, is now called the Mechanical Order Generation System, or MOGS.

    ITrac, on the other hand, was developed to better track inventory outside the mills—from production to processing to delivery.

    ITrac consists of three main elements: a piece that reports on the status of orders to U.S. Steel managers or its customers; a piece that tracks unfinished goods as they are moved between processors; and, a piece that handles messages over the electronic document network.

    The sweep of the order and inventory tracking systems, for a continuous flow manufacturing business like steel, "is astonishing," contends BCG's Shanahan, a consultant to U.S. Steel. "They can go from the shop, through their own production facilities, all the way through a third-party services center, to customers—and they can manage that entire supply chain through one integrated system.''

    Today, 38% of the company's orders for sheet metal and tin orders come in through the supply chain management system. Better forecasting means U.S. Steel supported $6.4 billion in revenue in 2001 with $870 million of inventory, compared to $946 million of inventory when it generated $6.1 billion in revenue in 2000.

    More simply put, U.S. Steel now keeps just 20 days' worth of inventory on hand to meet demand, where it kept 33 days' worth on hand in 2000. That means tens of millions of dollars saved every year.

    Hewing a Straightline

    Which led U.S. Steel to found Straightline.

    By early 2000, the management of U.S. Steel, which was then part of USX, was looking for ways to get its stock price up. The company brought in Arthur D. Little to help.

    The ADL consulting team, led by Shanahan, concluded that U.S. Steel's supply chain management system, which was boosting productivity, reducing costs, and bringing partners together via the electronic network, was saleable.

    The systems were "perfectly designed for steel." And since they were good at serving large customers, they likely could be adapted to help U.S. Steel serve smaller customers such as the Pate Company, a $5 million metal roofing products manufacturer, and Marconi Plc.'s Outside Plant and Power division, which makes telecom-equipment housing for the $9.4 billion communications company.

    These were customers who bought U.S. Steel products through service centers. Any U.S. Steel operation would have to offer superior services.

    But with features that would allow customers to know when products would be delivered almost as soon as they were ordered, the company thought it could build and maintain an advantage, through information delivery.

    "The way we look at ourselves, part of our strategic differentiation is the fact that we can manage information very effectively to help out customers," says Straightline Chief Executive Officer Robert Dryburgh.

    The iTrac inventory management system and the message network became integral parts of Straightline. Trudell's group also thought about using U.S. Steel's order management system, but realized Straightline needed a system more finely tuned for electronic commerce.

    Straightline brought on Web-based applications that would be hard for other service centers to match, such as online credit checking and order aggregation.

    For order aggregation, Straightline is using a product called SmartTrim from a company called Strategic Systems International. SmartTrim, first used in the paper goods industry, combines similar orders and then figures out means of moving products through the steel company's portfolio of processors in chains that eliminate as much waste and scrap as possible. SSI president Shoaib Abbasi said he expects Straightline to realize up to a 5% improvement in yields; that's the amount of product that can be produced from a given amount of original material. And, he notes, even cutting seemingly insignificant amounts of waste out of a production process can result in millions of dollars in savings over the course of a year for U.S. Steel.

    Looking for more millions from seemingly small savings, Straightline chose LiveCapital's DecisionExpress software to speed up authorization of credit so buyers could purchase steel promptly, when wanted. The customer gets quickness, while Steel's savings are in reducing uncollectible debts.

    But Straightline says its real power comes from the sum of the parts. Straightline says when it prices a product for a customer, it's able to take into account everything that will affect the price of that order. Factors such as determining whether the inventory of needed steel and processing capacity is available to fill orders lets Straightline quote extremely competitive prices in seconds. Other service centers sometimes need a day or two.

    The main systems were bolted together in about seven months. As a result, there were some hiccups.

    Synchronizing data was a chore, at first. In order to pull together a complete picture of orders, SSI's SmartTrim software needed to take a quick snapshot of data in Straightline's order management system. So a program was written in the Structured Query Language that allowed SmartTrim to pull data from tables in Oracle 8 databases. Only then could it create a view of what was going on with inventory.

    But, according to Abbasi, it took anywhere from 10 to 15 minutes to complete a tabulation and synch up the two databases. Straightline rebuilt the process, adding several new indices and queries, so the two databases can be synchronized every two to three minutes.

    There were other hurdles as well.

    Straightline applications run on Sun servers in U.S. Steel's data center. But U.S. Steel admits that it took a little while to learn how to manage a Sun data center. When Straightline came online, the company did not install monitors to track the performance of pages served up by its Sun machines. If a Web page hung up for some reason, the technology team did not know about it until a user called. U.S. Steel has since installed a performance monitor, Patrol, from BMC Software.

    But software idiosyncrasies weren't the only problem.

    When the business was conceived in early 2000, its potential customers were still enjoying the fruits of a decade-long cycle of growth in the U.S. economy. But by the time Straightline was ready to begin test operations, in July 2001, the technology-driven dot-com bubble had burst, the country was mired in recession, steel prices were tumbling to all-time lows, the company was racking up its worst losses in a decade and it was distracted by a breakup of the company, as sibling Marathon Oil was split off from steel operations. U.S. Steel fell from being part of the 43rd largest company in America, with $33 billion in revenue, to a stand-alone company that would rank only 287th in the Fortune 500 if it had been independent last year.

    Perhaps not the best time to launch a news sales initiative. Yet U.S. Steel plowed ahead.


    Customer or Competitor?

    The group that met in early 2000 narrowed in on the service center arena as the best opportunity to boost sales.

    No single company dominated the service center business. Three of the biggest players in the market include Ryerson Tull, which had 2001 revenue of $2.2 billion, Reliance Steel & Aluminum, which had revenue last year of $1.6 billion, and Metals USA, which reported 2001 sales of $1.5 billion.

    Few, if any, service centers were technology experts. U.S. Steel also knew, Shanahan says, "that the service center market was not a ferociously competitive, intense business."

    That smelled like opportunity. Service centers of all stripes in the steel business pulled in revenue of between $30 billion and $40 billion each year.

    But U.S. Steel does 15% of its business through service centers and was afraid it might alienate its 120 allies that served smaller customers.

    So, Straightline was set up as if it were an independent company. Straightline was free to establish its own offices outside of U.S. Steel and, within those offices, it could largely set its strategy and operations as it saw fit. This even included the recruiting of other steelmakers to supply material to Straightline customers. Straightline currently markets steel from U.S. Steel, Steel Dynamics and other manufacturers.

    Straightline offices are located about a half-mile from U.S. Steel's headquarters. But the distance between the two organizations in terms of business culture and psychology is extreme. Long halls and big executive offices fill the block-long glass and steel fortress complex of U.S. Steel, the tallest tower on the Pittsburgh skyline. Straightline's office-less headquarters, on the other hand, are tucked inside a three-story alabaster building nestled among scores of other nondescript buildings.

    In fact, there are only two private offices at Straightline, one for the chief executive officer, Dryburgh, an accountant turned steel and construction executive, and the other for the human resources manager. The walls are devoid of art, photographs and awards or achievements. "It sends a signal," Dryburgh says. "We want to be a low-cost provider."

    Outside Dryburgh's office, nestled among the cubicles, are gathering spots—couches, chairs and coffee tables—around which staffers talk, make plans, and solve problems. There's plenty of open space, a kitchen and even a foosball table for the employees when they take a break. It's almost as if Straightline aspires to be a 1990s dot-com market maker.

    And, in fact, some have made that comparison. "I think what U.S. Steel is doing with Straightline is innovative," says John Davis, vice president of purchasing, information technology and engineering at National Steel, the nation's No. 3 steelmaker. "But I'm not sure what returns on investment they're going to get. We can look at all the failed dot-coms and each one had a vision similar to Straightline."

    But Straightline says it has no illusions. "We're not a dot-com," Dryburgh says. "We look at ourselves as a distribution business."

    That's clear to U.S. Steel's distributors. Many centers are irate that U.S. Steel is now competing against them.

    "It's left a bad taste in our mouth," says John Applegate, a product manager at Viking Materials, a service center in Minneapolis that has been a U.S. Steel distributor.

    Bill O'Neal, CEO of O'Neal Steel, a Birmingham, Ala., service center, is a major customer of U.S. Steel and an outspoken critic of the Straightline launch. He thinks it's going to be difficult for U.S. Steel to be both a competitor and a supplier to the service center industry.

    But U.S. Steel says it has not lost any customers. And Applegate says there