Armco, Inc.: Midwestern Steel Division * In January 1991, management of the Kansas City with slightly over $1.7 billion in net sales, and operat- Works of Armco's Midwestern Steel Division began ing profits of $77 million. implementing a new performance measurement sys- Armco's Midwestern Steel Division generated tem. Bob Nenni, Director of Finance for the Midwest- $550 million in sales in 1990. (A division organization ern Steel Division, explained: chart is shown in Exhibit 1.) Within the division, the With our old system, our managers spent more Kansas City Works was by far the largest entity, ac- time explaining why changes in costs were caused counting for approximately $250 million in sales. Like by problems with our accounting system than they that of most of the firms in the U.S. steel industry, did fixing the problems. The new performance business at the Kansas City Works had declined signif- measurement system is designed to give us better icantly in the last decade. Employment was down from management focus on the things that are most im- 5,000 employees in 1980 to 1,000 in 1990.' The Works portant for them to worry about, carlier warning had recorded significant losses in the decade of the of problems, and improved commitment to 1980s, but it had been marginally profitable since achieve objectives. 1988. In the summer of 1991, the new system was still The Kansas City Works produced two primary being implemented and its design refined. But Bob products: grinding media and carbon wire rod. Grind- Nenni believed that the new system would be success- ing media were steel balls used for crushing ore in ful at the Kansas City Works, and he hoped that its use mining operations. Carbon wire rod was used to make would spread throughout Armco. shopping carts, bed springs, coat hangers, and other products. In 1990 the Kansas City Works sold 700,000 BACKGROUND OF ARMCO AND tons of steel: 200,000 tons of grinding media and THE KANSAS CITY WORKS 500,000 tons of rods. Armco was recognized as the Armco, Inc., was a producer of stainless, electrical, leading supplier of grinding media products in the U.S. and carbon steels and steel products. Through joint Armco's balls had proven themselves to be the most ventures the company also produced coated, high durable, and Armco received fewer customer complaints strength and low-carbon flat rolled steels and oil field about its balls than did its competitors. Carbon wire machinery and equipment. In 1990, Armco was the rods, on the other hand, were basically a commodity sixth largest steel manufacturer in the United States product. Armco's rod mill, which used relatively old Over the same period, Armco, Inc., decreased in size from * Copyright @ Professor Kenneth A. Merchant. 70,000 to 23,000 employees.Technology, was not cool competitive, so pods were moes section! Finally. the bars wang processed into not a profitable product. But the rods did genenite finished rods or balls. The rod mill shop worked with volume and helped cover some of the fixed costs of square cross-section bars. It malused the borg' diameter the plant. to belucen 4" and %" and coiled them into 20100-pound The Karees City Works was not a low cost manu bundles for shipment to customers. The rods were fur. faccover Its union labor costs in Kansas City were her reduced in size ( cold drawn to win ) al the cu. higher than those of some of its nonunion competingra towners' facilities for use in their products. The grinding particularly those located in the Southeastern LS, and media shop worked with the circular cross section bars non-L.S. locations, And the Works had an inefficient produced by the 19" mill II formed them into spheres plant infreedpicture because the plant was designed to using a roll-forming machine. Finished balls ranged in accommodate five times as many employees as were size from (" mo 5" in diameter. cun lilly working there. Instead of being efficiently laid out, the buildings alill being wed were spread across a 900-acre plant site. CRITICAL SUCCESS FACTORS Bocause of the plant's over disadvantage, the Work k IN THE WORKS manngers looked for ways in differentiate their prod- A. The Mew Shop ucia wod to develop new higher value products, and they had had some success in doing so. Each year ap- The mell shop, which included the ladle are fumace proximately 10 percent of the shipments of the Kansas raid the continuous caster, produced molten steel in City Works were of new higher value, high carbon 167-lon balches known as "heats." The shop's goal win lo rin three "turns" (shifts) a day, seven days a week, content products All salaried employees in the Works wire eligible 50 weeks a year, excluding the eight hours a week used for preventative maintenance. The other tap for cash incentive awards based on a performance evaluation made by their immediate superior and, ulti- weeks of the year ware used for expensive preventative mately, Rob Cushman, the division president. The in- maintenance and installation of new equipment. The mell shop could theoretically produce about 10 heads creative award potentials ranged from approximately 5 percent-30 percent of annual salary depending on week, but the best quarter it had ever achieved was an the individual's organization level. The performance average of 93 heawwink evaluations wise subjective but were based on, type- For a number of nasons, good performance in the cally these measures of performance applicable to the melt shop was critical to the performance of the position. For example, Rob Cushman described the Karexa City Works as a whole. First, the melt shop was criteria he used for evaluating the performance of the "boulenock" operation, to output from this phase Charlie Bradshaw, the Works Manager, as being bared of manufacturing process debamined the output of the plant as a whole. approximately one-third on plant safety, one-third on hard production numbers (particularly productivity Second, the melt shop costs accounted for nearly 40 percent of che total stoel conversion costs incurred and quality), and one-third on his evaluation of Char- lies "leadership" (Le. "Do I hear good things and see in the plant. The largest expenditures in the melt shop good things going on ?']. were for labor, production materials of various types and energy. Energy alone accounted for approximately 10 percent of the mel shop costs. "Works managers THE MANUFACTURING PROCESS were working toward computer control of energy, but AT THE KANSAS CITY WORKS in 1991 the melt shop manager still made most deci- sions about the heat used in the furnace, a major The manufacturing process wad for making both rods energy consumer. In 1983, Ance made an $8 million and grinding media included four basic steps. First. investment in a new ladle nice furnace that significantly scrap steel was melted in the ladile are furnaces, Second changed the melting furnace technology med in the the melied stral was poured into a continuous caster plant, the costs were declining as the malt shop mak that producial solid bars 30 feet in length with a 7" by T cross-section. Third, the 19" mill pressed the steel bars swears learned how best to use the new technology between two large cylindrical rollers to give them The dhiance lipa center lo canter of the two prewing mola eichix a square or chcular shape and with either ]" or 4" was 19 inchin. Henog the name for the process.Third, the quality of the raw steel prochased by the THE OLD PERFORMANCE mell shop wa ait important component in determining MEASUREMENT SYSTEM whether the finished products met the required specific The manufacturing areas of the Kansas City Works cations. Quality was affected by the grades of scrap were divided into five responsibility centers. melting. steel and nonmetallic materials used in the process. eating, the 19" mill, the red mill department, and the Nonmetallic materials were consumable items added grinding media department. Each responsibility center to batches to remove contruminants from the steel. was comprised of one of more cost centers Anmoo managers purchased a variety of pracdys of Before changes were made in 1991, the penfor- scrap steel and nonmetallic materials, and they used mances of the com center managers and their superiors different proportions of semp to nonmetallic materials in the plant were evaluated in terms of cost control and depending on the grades of scrap and nonmetallics safety. The key oded parkoumince measure was a sun- being used: lower grades of scrap typically contained mary measure called "Cost Above" which included the more contaminants Some of the production popcesses cost added perton of sted of each production stage and were standardred, with the addition of some now for the entire plant. Cost Above and the Meme that com metallics doun either by minemated equipment or by prised it were reported to the manubicturing managers production employees following standardized recipes on an Operating Sitistics Report that was produced on Other processes, however, required the manufacturing approximately the 15th day following each mo thend manager And his technical supervisors to exercise The Operating Statistics Report provided a flys- judgment. year history, monthly and year-to date actuals, and B. Rolling and Finishing monthly and year-in-date objectives and variances from objectives for each of the factors that determined Personnel in the Rolling, and Finishing fren, were total Cost Above for each cost center. Exhubit 2 shows asked in make parts to specification while controlling a portion of the Operating Statistics Report for one Fields and costs. Customer specifications for codes ugu- oost center-the $2 Mell Shop' (The entire report, ally contained physical property requirements, such is printed on five computes pages, included detailed in- for ductility and elasticity One specification for bells formation about 46 separate expense categorise ) The required a teo-story doop test. If the test ball cracked report also gave cost per net ton [SINT) for many of into two parts on impact being dropped from tum the cost categories. The Tolil Cam Above NT is thewn stories, then the product was rejected on quality in the next to last column of page 3 of Exhibit 2. grounds. In addition, the lives of the balls were tested The Operating Statistics Reports used the same ac- in Armco's customers' actual grinding operickna counting information that was used for financial re- Those tests had shown that Armco's balls were more porting and inventory valuation purposes, so the fig than competitive, they lasted up in 15 percent longer area included allocations of indirect manufacturing than did its closest competitor's balls. The colling areas costs. For example, to provide smoother cost patterns, were heavily capital intensive. Simificant oasis in the the charge for the two-week plant maintenance shut- finishing areas were for labor, energy, maintenance, down was spread over the 30 weeks of operations and yield losses. These costs, which incheaded labor and material, were C. Maintenance shown on the Operating Statistics Report is"S Onder" costs. Maintenance was also an important determina! of The operating managers had become accustomed success in the Kansas City Works. The goal of mainle- In the Operating Statistics Report, and in genial they nance was to maximize equipment up-time while con- liked it. For example, Gary Downey, the Melting Op- trolling maintenance expenditures. Organizationally, orations manager, said that he looked at 95 percent of the maintenance activities were divided into three the information presented in the report, although he groups. Teams of electrical and mechanical mainte- acknowledged that some of the items were quite small mance employees were assigned in each manufacturing in dollar value, Paul Phillips, the Rolling and Finishing cost center A third group operated a centralized main- manager, liked leaving the monthly and annual trends tenance shop. The cost of maintenance was significant, as approximately 40 percent of the TO0 hourly ci- The #1 Melt Shop contained obsolete equipment and was playees in the plant wise maintenance workers no longer used.and the information comparing acnul costs with ob reports, and product cost reports. One Bysam jectives. Paul felt that the Operating Statistics report can't do all these brings well. was "the minimum amount of detail nosesguy" He would have prefernal to have the Operating Smigtics The new system was designed both as a means of Report on a weekly basis and in his hands on Monday providing middle- and lower level manages with a morning because, for example, "If we see that fuel greater understanding of how their actions related boy Consumption is unusually high, we can go and look for the implementation of the division's business strate- the christ ." gias and as an improved method for managers at all The accounting department eles provided other re- levels to assess the extent to which the desired results ports showing the detail behind the figures for some of were being achieved. The vision and goals of the orga- the cost elements on request. For example, one report nization were to be translated into key success factors showed the cost for nowmetallic materials broken which would be disaggregated into department and in- down by the specific materials used. dividual objectives that would be compared with men- sures of noanal results. The basic philosophy is illus- crated in chart Corn in Exhibit 3. THE NEW PERFORMANCE MEASUREMENT SYSTEM EXHIBIT 3 Arica, Inc.: Midwestern Steel Division Vision Management Proces A. The Goals of the New Syntew Bob Heuni, the director of finance, had been working on the performant mecement system since 1989. but due to staff constraints be had been unable to de- sign and implement the new system while keeping the old system going. On November 1, 1990, Rob Cushmin was appointed as president of the Midwest- arn Steel Division, and Rob sponsored the implement Oub tation of a new performanos measurement system. He allowed Bob Nenni to discontinue production of the Operating Statistics Report in January 1991 in order to implement the now system. Rob Cushman obearval The old symem wasn't working People were rely ing on something that was not adequate.. . Enough companies are ming good parfor- males measurements is building blocks to excel- Cupagiant lense. I don't want to go against the grain. I wank In give my managers the information they need. And I want to have good measures that tell us how we've done. I'm not using the partormanos as a Indirkkal threat. I'm trying to make it fun so that when we determine we've done well we can celebrate our If this planit does everything right, we should be able in make $30 million per year. But we're not doing it. This system is part of a spirit of change that has to happen. We will give people more responsibility . . . and more latitude to fail. Bob Nomi added: The cast part of our old performance measure- ment spalein we built for accountants. It was de- Desired action signed to produces financial statements, operatingRob Cushman and Boh Nenal thought that the new specification, and percentage ou-lime shipment system promised to major improvements. First, the Spending was the accumulation of all expenses in- new system was designed so that managers would curved by the people reporting directly to a manage focus on the few key objectives that largaly determine The maintenance performance measures had not yet the success of the Eman City Works and not get in- been clearly defined by the middle of 1991, but main volved in the detail until a problem existed. As Doh lonance Indoor cost and material cost were being moga. Hemi observed: sured. Performance measure 7 chrough 10 were When marcigers get too much data, they can can- plant-wide (not cost canker) measure Cash flow was ily get unfocused. The new sysdent will cause measured monthly for the plant. Product mix was the them to focus on the 3 or & things that cause percentage of high carbon products sold compared to 89 percent of the spots, not the 40 that once low carbon. Inventory days on hand was tracked 100 percent. monthly Accountability for inventory performance was shared among plant purchasing mennigers man- Second, the new system was designed to provide an factoring managers, and commercial managers. Sales improved basis for evaluating operating managers and price minus net metal, a memeure of value added, was manufacturing supervisors. The systemn wookd include backed monthly. a balanced set of performance measures, including The design group discussed the couponcals of quality, schedule achievement, and safety, in addition such performance area and the ways in which each to canals, And che east reports would be improved be- measure could be disapprovaled to guide performance cause they would include only those costs deemed at lower management levels For example, the cascad. controllable by each individual operating manager. ing of goals relating to Total Quality is illustrated in They would not be distorted by volume changes as in Exhibit 4. Total quality at the Works level was affected the old system. by the proportion of products meeting customer speci- B. The Design of the New System fixations, yiekka, and percentage of shipments made The new system design process began early in 1950. on-time shipments, and each of these indicators could Bob Cushman, Charlic Bradshaw, Bob Nenm. Gil be disaggregated further. The intent ways to mezaire Smith (commercial director), and others defined 10 bey each of these areas of performance at the lowest rule- performance measures for the Kansas City Works: want level of the organization. One of the most significant changes was the climi- 1. Heats per week nation of Coat Above measure. Production managers 2. Tons per man hour were no longer held nocountable for all costs incurred 3. Disabling injury index in or allocated to their respective locas so, in effect, 4. Total quality index they were no longer cost center managers, The cost 5, Spending detail in the new performance reports was reduced considerably. In the new system, the only cost figure 6. Maintenance performance on which managers were evaluated was the spending 7. Cash flow by the canplayers in their organizations. For example, 3. Product mix in January 1991 spending on maintenance in the mask 9. Inventory days on hand shop was $300,000, but only 530,000 of this amount 10. Sales price minus cost of net metal was spent by people reporting in Gary Downey's or- ganization. Thus Gary's report included only the Performance measure I. heats per week, was only $30 000 figure. The other $270,000 was reported in relevant to the melt shop. However, since the malt other managers, particularly those of maintenance shop with the bottleneck operation, heats per week was foid purchasing I cribed measure for the Works as a whole. Measures 2 through 6 were applicable to all manufacturing C The implementation Process areas. Tons per man hour was a productivity measure. On January 1, 1991, Bob Newni discontinued the The disabling injury index was a safety measure. Operating Statistics Report system. He believed that The total quality index was the product of three mea- the new system would have had no chance if the sures: physical yield, percentage of product meeting "minapara kept using the old data and never seriouslyEXHIBIT 4 Armco, Inc : Midwestern Steel Division Cascading of Total Quality Goals puleg Hering Sounded THI arad wan considered improvements that could be made." The reports. This change was made to give the operating accounting department began the process of produce managers a number that they could compare to their ing new sets of reports As the entire mask could not budgeted spending bargers which had been prepains! be accomplished immediately, they focused their at- using the old welsurement philosophy Starting in tention on producing some piled reports for a'sbay 1972, however, they promised that the reports would of the memoires. They focused first on heats pix week. reflect only the new cost performance philosophy. By tons per man hour, physical yield (a component of then the performance targets would be set using that the total quality index), and spending, Exhibit 5 gives sam philosophy. an example of a report for the Melting Operations In June 1991, Bob Nenni reflected on nine months Manager. since the design meetings began He was convinced The operating managers" initial reaction to the sam- that the company win on the right track even though ple reports they were given was dissatisfaction The some of the managers were uncomfortable with the early reports did not provide the line-jum expense new syriem. And he knew that the delays in the imple- detail to which they had become aceretomed, In sadi- miamiation process had frustrated both the information tion, they ware no longer given Cost Aboy infor- Teams and his accounting raff He noted mation, so they asked, "Where are my spending numbers?" In early April 1991, Charlie Bradshaw told We're trying in change the way the manager Bob Nenad, "I've received nothing of me from your chink. The new system is not yet part of their man- department since you discontinued the old reports." Mlity. Changing mindseta is wtiminely more in- Another manager complained, "It almost seems like porbut and more challenging than the technical the operating managers finally undentood the old re- job of producing the reports. But we in accounting feel we will now be port, so they decided to change it." In late April, the accounting group backed off their more useful to the organization. We were spend ing 60 percent of our time in nosounding on the initial implementation plan. They started to provide mom-value-added chan's of inventory valuition spending numbers for the entire cost center in addi- for firsocial reporting purposes. We have now cian to the spending initiated by a manager's direct seduced that to 20 percent.REMAINING ISSUES to individual performance evaluations. In other words, In 1991, two related performance evaluation/incentive how much of total compensation should be provided in issues aroge in discussion One was an issue about fixed salary, and how much should be paid only to how to evaluate managers' performances in situations those who were good at getting things done and done where the numbers were distorted by uncontrollable well? factors. For example, early in 1991 the melt shop suffered two transformer failures, apparently because Questions of fluctuations in the line voltages provided by the local utility, Kansas City Power and Light. Such fail- L. What was wrong with the old system? (Hint Study ures had happened nearly every year, but shop man- Exhibit 2 carefully and figure out what the columns agers had recently upgraded some of their electrical tell you, individually and in total. ) switches to try to eliminate the problem. Nonetheless, the failures occurred again, and by April, Gary 2. If the old system was so bad, why did the operating Downey knew that his goal to average 101 heats per managers seem to like it? week was impossible. The failure of the melt shop to 3. Evaluate the new system and the way in which it achieve its plan would mean that the Kansas City was being implemented. What changes would you Works as a whole would not be able to achieve its recommend, if any? Why? plans for 1991. Rob Cushman knew that at the end of 4. What are (were) these systems designed to do? the year he would have to decide whether or not to let Would you consider either (or both) of them cost this, and perhaps other similar occurrences, affect the accounting systems? evaluations of his operating managers. 5. What should Rob Cushman do about the two items The second was an issue as to whether to increase described in the Remaining Issues section of the the production of total compensation that was linked case