Fluent Essays

9-699-133 R E V : M A R C H 1 3 , 2 0 0 3 ________________________________________________________________________________________________________________ Professor Andrew P. McAfee prepared this case. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 1999 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or go to http://www.hbsp.harvard.edu. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without the permission of Harvard Business School. A N D R E W P . M C A F E E Rich-Con Steel On a Friday afternoon in January of 1997, Marty Sawyer, president of Rich-Con Steel, shut the door on the chaos outside her office, sat down, and tried to gather her thoughts. She had a crucial decision to make—should her company continue to use its new information system? It had only been in place for a month, but during that time the company had almost completely lost its ability to carry out its most fundamental activities: ordering and receiving raw material, tracking inventory, filling orders for customers, and billing them. If it were obvious that the new system was fatally flawed, her decision would be easy: she would immediately cease to use it and go back to the previous information system, even though it was antiquated and primitive. Sawyer was considering this because she had noticed several areas where the new software did not appear to be working properly. It was not clear, however, that the system itself was solely at fault. The company had not implemented a new information technology for over twenty years, so Sawyer and her team could have made inappropriate decisions, or overlooked important ones, when setting up their new environment. Rich-Cons business was surprisingly complex, so it would not be difficult to misconfigure the software used to support it. In many cases, it would be difficult to distinguish a misconfigured system from a malfunctioning one. One other consideration complicated her decision. Whether or not the implementation had gone well, the companys employees could be using the new system incorrectly, or not at all. Most of them had never before worked with a similar system, or even with a PC. Given this, it was likely that they were making many mistakes. Were these mistakes responsible for Rich-Cons current difficulties? If so, could they be corrected over time? Sawyer was not even sure that most people were still using the system; there was already a body of workarounds in place to get work done without relying on the new technology. Should these be encouraged, or stopped immediately? Her father, Sam Sawyer, who had preceded her as president, knocked and entered. So, what are we going to do? he asked. Marty knew exactly what he was talking about, but did not know how to answer him. D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 699-133 Rich-Con Steel 2 Figure A: Rich-Con logo Company History Hardware Wholesaler Richards & Conover Steel Company prided itself on being the oldest business in Kansas City. It traced its founding to 1857, when John F. Richards arrived from St. Louis with the goal of selling hardware to settlers in the West. The company was well established as a wholesaler by the end of the nineteenth century, and exploited this recognition by affixing a “Rich-Con” logo to many of its products (see Figure A for a modern version of this logo). Rich-Con was focused almost entirely on distributing hardware, and did no manufacturing or processing beyond cutting sheet metal. The company prospered through the 1920s, suffered during the depression and World War II (when supplies of any kind were hard to come by) and profited from the immediate post-war boost in consumer demand. However, the rise of vertically integrated hardware chains in the 1950s sent the independent wholesale hardware business into sharp decline. Sam Sawyer, a great-grandson of John Richard who was president of Rich-Con during this period, took three steps to maintain the company. First, he liquidated the hardware business during 1958 and 1959, eventually selling all inventory. At the same time, he used proceeds from the liquidation to buy out almost all other shareholders, who were happy to receive some value from a firm with no evident future. Finally, he decided to concentrate on Rich-Cons small steel service center, which at the time accounted for less than 15\% of revenues. As Sam Sawyer explained it: We had decided that we wanted to stay in the steel business because it’s a direct distribution. You’re not selling to a dealer who is again going to re-sell it. We’re selling direct to the users— fabricators and manufacturers. There would always be a good little niche for a company such as us in there. Steel Service Center As a service center, Rich-Con held raw material inventory—in the form of steel sheets, plates, beams, bars, pipe, and tubing—from a number of producers.1 [See Exhibit 1 for examples of these products.] Customers, including fabricators and equipment manufacturers, ordered combinations of these materials from Rich-Con instead of dealing with several separate steel makers. In addition to acting as a consolidator, Rich-Con also processed pieces of steel for its customers, cutting and bending them as required. As a service center, Rich-Con grew steadily and formed close relations with a number of customers in the area. The company developed a reputation for excellent service and responsiveness. Rich-Con also became known for its thrift. Sawyer bought much of the firms shop floor machinery at 1 Rich-Con did not stock or process steel coils.D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 Rich-Con Steel 699-133 3 auctions held after other service centers went out of business. Once in place, a piece of equipment was typically used as long as it was serviceable. In the 1970s, Sawyer and Rich-Con decided to develop particularly deep expertise in steel rolling, or bending beams into desired configurations. Margins for rolling were comparatively high, and there were a few other firms in the area with advanced rolling capabilities. In 1996, the company purchased a number of rolling machines, including the Roundo, a press that was capable of bending into a circle the largest steel beams made (see Exhibit 2 for pictures of the Roundo and Rich-Cons other rolling equipment). The Roundo was the largest machine of its kind in the world and, at one million dollars, the most expensive piece of fabrication equipment owned by Rich-Con. By the mid 1990s, the privately held firm employed 140 people and had revenues that Sawyer gave as between 25 and 50 million dollars. It also had a new leader. In 1994, Marty Sawyer, Sams daughter, became president of the company. She had come to Rich-Con in the mid 1980s, and had worked in purchasing, marketing, and operations. Information Technology The Legacy Environment Marty Sawyer had also worked on the companys information systems, and strongly felt that they were outdated and in need of replacement. She considered replacing this “legacy environment” to be her major challenge as Rich-Cons president. The companys thrift had extended to its information technology. In 1994, financial and inventory management software ran on two separate IBM System Three computers. The System Three was introduced in 1969, and relied on keypunch operation for all data input. The software running on these computers had been largely written over the years by Rich-Con.2 This software did a reliable job of keeping track of material, and of updating the general ledger as inventory balances changed. However, these systems did not have the ability to keep track of orders. Like many companies, Rich-Con used customer orders to record required materials and processing steps, purchase orders to replenish supplies and pay suppliers, and work orders to send instructions to the warehouse and the shop floor. Unlike most companies, however, Rich-Con in 1994 was still writing all of these orders by hand. As Credit Manager Donna Ehlers explained it: Orders were hand-written from beginning to billing. At the time the salesmen took the order, it was hand-written on a work order-type form. That was a five-part form and it was broken down for order fillings. That form traveled around the company from the time the order was taken until it was billed. And it was billed in a batch keypunch environment. Marty Sawyer considered processes like this inefficient and unreliable, but she also had a more fundamental concern. It was virtually impossible to obtain useful historical information about the companys activities. The System Three did not include a database, so each data extract required custom programming that could take months. The only alternative—going through stacks of paper orders each time information was needed—was also laborious. As a result, there was no effective 2 The programming language used by the System Three was an obscure one, and in 1994 Marty Sawyer did not know of anyone outside the company who could still write code in it.D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 699-133 Rich-Con Steel 4 way for anyone at Rich-Con to analyze trends over time in sales, inventory usage, scrap rates, or other quantities of interest. Keeping the System Three computers up and running was also a challenge. Replacement disk drives, terminals, and other hardware for a 1969 computer were very difficult to find by 1994, and Rich-Con eventually resorted to scavenging. Marty Sawyer recalled that: We started getting free equipment 10 years before we stopped using the system. Im sure we were the last company in the country to be using that machine. Sam Sawyers memories of finding replacement hardware were also vivid: They just weren’t making parts anymore. So, the only way we could get parts with this was to cannibalize old machines that we could pick up. And we had a warehouse full of them across the street. And we were picking these machines up, literally, for nothing. People would give them to you if you would haul them away. So, when we finally scrapped that over there, somebody estimated we had at original cost something like $32 million worth of computers sitting over there. We had 70 or 80 machines over there, just to keep ours working. And we sold them for scrap for $625. These spare parts were crucial. Marty Sawyer estimated that in the mid-90s the computers would, on average, “go down” once a month for several hours, or even days, until the problem could be diagnosed and replacement hardware installed. In-house Development In the early 90s, Rich-Con had attempted to write its own software to replace the System Threes and the applications they ran. The company had bought an accounting package and AT&T hardware running the Unix operating system, and began making extensive modifications to the package. Marty Sawyer had been in charge of this effort, and realized by the end of 1993 that it was not going to be successful. She recalled: It just wasnt working. It was just taking forever and we were still nowhere close. We were working on the order entry and there was still so much to do and so many bugs and I just knew this was not going to work. I never would have had the guts to turn it on, I dont think. And I just wanted to be able to have a system that would work and I wouldnt have to worry about it. So, we just decided to pull the plug on it and go out and buy the integrated package. Selecting a Standard Solution In 1995, Marty Sawyer and Rich-Con began to look for an off-the-shelf software package to help them run their business. Over a period of several months, she investigated approximately half a dozen packages, all of which ran on the Unix operating system and were written for the metal industry. The company stuck with Unix so that it could continue to use the AT&T hardware purchased earlier. Sawyer felt she had to use software written specifically for the metal industry because more general logistics and distributions systems, despite their attractive features, could not handle the unique requirements of a steel service center like Rich-Con. These requirements included: ! Multiple units of measure. The same piece of steel stock could be referred to and priced in several different ways—by length, by weight, or by quantity. As SawyerD o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 Rich-Con Steel 699-133 5 explained it, Some people want it priced per 100 foot. Some people want it per foot. Some people want it per piece. Some people want it as a lot. Some people want it per 100 weight. Some people want it per pound. And sometimes they want the quantity in one unit of measure, but the price in another unit of measure, like Im going to buy 10,000 pounds, so give me a price per foot. Software written for the metal industry typically had this ability, but more general systems did not. Sawyer felt that this was an essential function of any package Rich-Con purchased. ! Remnants. Very often, customers wanted only a section of a steel sheet or beam. After cutting off the required section, the rest of the stock would be returned to inventory. Rich-Con wanted to be able to track these three-dimensional remnants, or rems, on a piece-by-piece basis so that they could be used to satisfy other orders. ! Certifications. Many types of steel were accompanied by certifications, or certs, of their mechanical and chemical properties provided by the manufacturer. Customers often wanted these certs delivered along with the materials. Rich-Con wanted to be able to store certs in their information system, and print and deliver them as part of an orders paperwork. When investigating candidate systems, Sawyer asked questions to determine if a package could meet these requirements. She also asked more general questions, on issues such as how an application calculated when it was time to order new materials, and whether it could handle intra- company inventory transfers. The systems that looked promising were demonstrated at Rich-Con, and Marty Sawyer made several visits to see them in operation at other companies. Sawyer found that no package she investigated could completely satisfy all requirements. However, she found on commercially available application that, in her opinion, came close. This application came from an established vendor, ran on Unix, and satisfied many of the unique requirements of the steel industry. In addition, she was impressed with its order entry capabilities. She visited several sites where it was installed, and so became convinced that the vendor was not just selling vaporware. The vendor was willing to sell its software without an accompanying maintenance contract, which Sawyer felt would save money. The vendor also indicated that it would be able to accommodate any requests to modify the software. Rich-Con selected its application to completely replace its legacy systems, and began implementation activities in late 1995. Rich-Con did not hire an external consultant to help with the implementation project, relying instead on the vendor’s implementation support resources. System Implementation The project consumed a great deal of Sawyers attention even though it was, in many ways, an unexceptional information system implementation. Rich-Con did not attempt to write a great deal of ”custom code,” or modifications to the standard software. Instead, Sawyers time was taken up understanding the softwares capabilities, configuring it to match the way Rich-Con did business, populating its databases with information on existing inventory, customers, and suppliers, and overseeing training. Sawyer felt that training was particularly important because many users of the new software had no experience with modern information systems, or even PCs. They would thus require a great deal of preparation on how to use the new system and the hardware it ran on, and on how their jobs would change once it was in place. However, Sawyer was unhappy with the training resourcesD o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 699-133 Rich-Con Steel 6 provided by the software vendor. She commented, We had people come in and do training. It just never really took on a life of its own. Sawyer was also surprised that the vendor did not seem to have a standard methodology for implementing its system at a customer. She expected that the software vendors employees would ask an established set of questions about how Rich-Con did business, and use the answers to configure the software, recommend business process changes, or both. Instead, she felt that during the implementation personnel from the vendor would show up and wait to be asked questions about how to use the software. Sawyer did not feel that this was an effective approach, but attributed it to the software vendors emphasis on product, rather than implementation, strength. Because the vendor did not appear to be guiding the implementation, and because of her expertise with Rich-Cons existing information systems and business processes, Marty Sawyer became deeply involved with the project at the same time she was settling into her role as company president. By the end of the implementation, she estimated that she was spending 75\% of her time on it. This meant that she had to rely heavily on other people, some of them newly hired, to manage much of the companys operations. This proved to be damaging. As Sawyer recalled: As we got closer to turning the system on, I was spending more and more of my time overseeing the project. Unfortunately, with a new management team and very little computer expertise within the company, it fell to me to get the system live. Additionally, a lot of other distracting issues were going on in 1996. We had an ongoing union problem, and we were setting up the rolling division to house the new Roundo we had recently purchased. Sales were dropping, but we did not really know what at what level without a lot of hand analysis. The only performance analysis we had at the time was a listing of customers and their sales for the quarter. I felt very strongly at the time that to identify the problems and to create a plan to improve the company performance we needed information, which meant a new computer system. The real problem, though, was that the new management team was not performing and these other issues in the business were disguising that fact. Going Live Finally, in December of 1996, Rich-Con went live with the new system, and began to use it to track orders from customers, purchases of raw materials, and inventory. Sawyer explained the timing for this event: I didnt feel like we were completely ready but it was kind of like if we dont turn it on now, well never get it on. December is a slow month. And I was really scared about this old hardware. I needed to get off of it and I thought this software would work. Also, we knew we would take a complete physical inventory a month later. I said, “Well if we screw it up, were going to do an inventory a month later.” Problems with the New System Symptoms Very soon after the system went live, Marty Sawyer begin to get indications that all was not well: In the first week I was working until after midnight trying to get billing through the system. The Sales department was, I thought, doing okay. After the week was over we knew we hadD o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 Rich-Con Steel 699-133 7 some major problems because we couldnt tie the system; we couldnt get billing records to reconcile. What I didnt know was that the system had also set off a bomb on our shop floor. About two weeks in, we found out the shop wasnt getting their paperwork, so to take care of customers they were shipping on hand paperwork. Probably not half of our shipments over the first few weeks went out on system-generated paperwork. The rest went into a filing drawer with no reconciliation to the computer. My first thought was to stop people from going around the system, but they had to because the paperwork wouldnt come out. We still had to ship. Customers called screaming at us. We said, “We dont have the order!” It seemed that no matter what we did, a lot of the shop paperwork wouldnt come out. It was a train wreck. The problem was by the time we realized we had this problem we had a month of hand paperwork, and by then it was like, “How do we clean up this mess?” By the end of the year, it was a disaster. We had to take year-end inventory and we were screwing up everything: double shipping, double invoicing, not invoicing, not shipping. Because we were going around the system we were shipping on hand paperwork, then when the paperwork finally came out, we would ship it again. Or we would think that wed lost an order, and the sales desk would enter it again. It was a vicious cycle. After the first 30 days, it was unbelievably bad. It was just unimaginable. Sam Sawyer, who was still actively involved in the company, saw many of the same problems: We couldnt get our orders out. The customers were getting wrong products. They werent getting their billing properly. We couldnt get it there on time. Back-orders were getting lost. The sales staff couldnt rely on the inventory, and they would sell something, send you out to the plant, and it wasnt there. Or vice versa. Youd have a bunch of remnants out there that the sales department didnt know we had. So, there was just dead inventory sitting there. Or wed buy more, not knowing that we already had the inventory--already had plenty of it on hand. I mean, those are typical of the kind of things that were going on. And on the other end, customer invoices couldnt be matched with the initial paperwork. I dont know how much steel went out and never got billed. Causes Marty Sawyer was extremely worried about all of these problems, but she was even more troubled because she couldnt be sure what was causing them. She didnt know whether Rich-Cons constant difficulties were due to faults in the software, configuration mistakes made during implementation, misuse or non-use of the system, or a combination of these three. She noted: We thought, “Well, the employees are sidestepping the system. If they stopped going around the system, it would work,” as opposed to, “if they had gone on the system, it wouldnt have worked anyway.” Other people in the company were also having trouble identifying the source of Rich-Cons troubles. Depending on their jobs and their interactions with the new system, they saw different problems. Darrell Russell, the Vice President of Operations, had worked at several other companies D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 699-133 Rich-Con Steel 8 before coming to Rich-Con, and had extensive experience with other information systems. He felt that some of the logic in the system was flawed: With an MRP system, if youre manufacturing, you need to purchase all of the parts and pieces and have them available in inventory before you send your paperwork to the shop floor, to actually assemble the order. So, it doesnt make any sense to send out all of your paperwork until you have all of the pieces there to assemble it. But were not really a manufacturer, were a distributor, and we could, out of ten items, ship eight. And our customers would allow us to ship eight, and back-order the other two. A pretty standard way of doing business. But we couldnt the system to let us do that—we couldnt get the first eight out because it held everything because the other two werent here. The system would not recognize the fact that you could ship before you have all the items. Donna Ehlers, the Credit Manager, was a longtime Rich-Con employee who was comfortable with the information systems and procedures in place within her department prior to the implementation. She saw that the new system significantly affected the way people did their jobs, and she felt that many had become overwhelmed by the changes: What I saw were the people problems. We were not prepared at all. We were not trained. We were not mentally prepared and we were not emotionally prepared as a company when we went to the new system. And so, when we finally plugged in, you know, just, we were in shock, because nothing went well. Not order entry, not operations, not accounts receivable, not anything. No one had ever really been involved in implementing an online order entry system, or using one. The Decision By January of 1997, Rich-Con was in the middle of a crisis. Looking back on that time, Sam Sawyer recalled: That was the most difficult time since Ive been with the company. There was no other crisis in the economy or in the operations of the place that ranks with it in terms of how bad it hurt us. And the economy was pretty good, so we could have been selling a lot of steel. We should have been coining money, and instead we were losing it. Sam and Marty Sawyer had to decide what to do. Should Rich-Con continue to use the new system and attempt to identify what was causing the problems? What if they couldnt, and nothing got better? The company simply could not afford to operate for long in its current state of chaos: it would go bankrupt. So should they “pull the plug” immediately, and go back to their previous way of doing business? If so, what would happen when the System Three finally gave out? And even if it continued to run, how much was Rich-Con forsaking by not having better information about the companys operations? How large a handicap were their current primitive information systems? How large a handicap could they become? D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 Rich-Con Steel 699-133 9 Exhibit 1: Examples of steel products stocked by Rich-Con Concrete reinforcement bars (“Re-bar”) House columns Beams D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 699-133 Rich-Con Steel 10 Exhibit 2: Steel Rolling at Rich-Con The Roundo machine Other rolling equipment D o N ot C op y or P os t This document is authorized for educator review use only by Ron Davis, University of North Alabama until Jan 2022. Copying or posting is an infringement of copyright. [email protected] or 617.783.7860 RICH-CON CASE STUDY QUESTIONS Summary: One month after going live with a new information system, Rich-Con Steel is deep in a crisis. The new system was intended to give the company basic visibility and control over its business processes, but instead visibility and control appear to have evaporated. 1.a. What are three aspects/areas in the case where Rich-Con performed well? b. What are three aspects/areas in the case where Rich-Con performed poorly? 2. What do you think about Rich-Con’s approach in designing and implementing a new system? 3. How did the new system cause the chaos within the company? 4. How would you judge Marty Sawyer’s performance as president of Rich- Con? 5. If you were Marty Sawyer, what decision would you make at the end of the case about resolving the crisis that was caused by the new system?