PAPER PRODUCTION FOR FORNAX AT THE PLUTO MILL
E d Michaels, the recently appointed director of quality assurance at Great Western Papelco's Pluto Mill, did not much like what he had just heard at the mill's morning meeting. This meeting, a longstanding Pluto Mill tradition, was attended by the mill manager, his direct reports, and most of their direct reports-typically 17 persons. It was held each day at the start of the first shift. The nominal purpose of the morning meeting was general communication and planning, but, in fact, most of the discussion usually focused on things that had gone wrong during the previous day, firefighting efforts under way, and the like. The wrap-up of the morning meeting was a report on the previous day's production and shipment figures. An enormous chart plotting daily tons shipped dominated the conference room wall, and before the meeting started, the production manager was obliged to have yesterday's figures posted. The mill manager was obliged to telephone these figures personally to the Salt Lake City headquarters as soon as the meeting ended-usually to the White Paper Group vice president himself. This morning there had been a special topic-the upcoming run of Fornax reproduction paper, the first
to happen at the Pluto Mill in more than 6 weeks. The schedule called for 2400 tons of paper with production to start on Monday morning, if the paper makers could get the paper machine tuned up in time. The group vice president of White Papers, George Philliston, was leaning heavily on the Pluto mill manager, Rich Johnson, not to repeat the disastrous customer rejection that had occurred on the last run of Fornax paper. At the morning meeting, Ed and the Pluto Mill quality improvement lead team had been charged to help "do it right this time," and he was worried. This was a crucial time for Great Western and the Pluto Mill. The Great Western Papelco Company, a large, fully integrated forest products company, was just 8 months into a massive, company-wide total quality management effort that had been named "Quality Is Everything." To kick off this effort, all mill management had been through a 2½-day quality awareness training, new quality posters were on the walls of the company conference rooms, and Ed Michaels had been the first person appointed to hold the new position of mill director of quality assurance. In this job, Ed reported directly to Rich Johnson, the mill manager. Ed had, of course, gone through the 2½ day course and, in addition, had just returned from a 1-week training course on statistical methods for quality improvement, run by nearby Erehwon University. Many new and very technical ideas had been covered in those 4½ days at Erehwon. Ed came back to Pluto with a binder of notes that was fully 3 inches thick and a laptop with statistical process control (SPC) software. Pareto charts, control charts, Cp ratios, Ishikawa diagrams, and standard deviations filled his mind. In 3 months Ed would be going back for a second week of training that promised to include regression analysis, experimental design, and hypothesis testing. (It had been 17 years since his last math course in college.) Some of Ed's colleagues in the mill, and indeed Rich Johnson, the mill manager, expected this new methodology of SPC to be the silver bullet that would magically solve Pluto's serious quality problems. Ed wryly noted that Rich, a crafty old-timer, had scheduled himself to be out of town during the original quality awareness training. "Ed, I'm counting on you to lead this quality effort," were his parting words. Indeed, there were many serious quality problems. A task force consisting of outside consultants had, at the outset of the Quality Is Everything program, estimated that quality nonconformance costs at the Pluto Mill were an incredible 19% of revenues. Fully 70% of all plant overtime had been traced to specific quality problems, many shipments of even commodity-grade product had nonconformances on one or more specifications, Monday morning absenteeism was 11%, and the list goes on. Members of the senior management team of Great Western Papelco were already talking enthusiastically about the positive impact on corporate profits if even half of these qualityrelated costs could be reduced through SPC. But Ed Michaels, no newcomer himself to Great Western Papelco or to the Pluto Mill, had seen a series of quality programs introduced over the years with great fanfare, but little follow-up or long-term effect. There had been a Zero Defects program in the late 1960s, then a Quality Circle initiative in 1978. The latest program had been an Overhead Value Analysis in 1982. Though each program had brought initial benefits, none had delivered on its long-term promises and, in fact, they left most employees cynical or worse. "At least half the mill probably thinks Quality Is Everything will be another fiasco," mused Ed, "and this time I'm the point man." Ed knew that senior management had discussed the shortcomings of the past programs and was stating forcefully that Quality Is Everything would be different, but he wondered how real their commitment would be as the paper market continued to soften, and both prices and sales volumes dropped. The Pluto Mill, one of seven paper mills in the White Paper Group of Great Western Papelco, manufactures a variety of white papers for high-speed printing and reprographic applications. The Pluto Mill, indeed Great Western itself, has a very fragile relationship with the Fornax Corporation, a major producer of reprographic equipment. Pluto produces 8½ × 11 inch paper for the Fornax Corporation, which is sold under the Fornax Company label, and is used in the Fornax Fourth-Generation SuperDuper Hi-Speed Laser-Phasor Publishing System. Recently, Pluto made a large run of "Fornax" paper that, according to the tests done at the paper mill, met Fornax quality specifications. Nevertheless, the shipment was rejected by the customer when it was inspected upon receipt at the Fornax warehouse. The rejection was based on physical testing done according to a military standard statistical sampling plan. The tests were done at Fornax's own lab by the Fornax Quality Assurance Department. This situation caused a crisis at the Pluto Mill, as the return of the entire Fornax shipment cost Great Western Papelco well over $100,000. (What do you do with over 800,000 reams of paper in Fornax wrapping and labels?) Of course, Fornax itself lost $25,000 and more than a little customer goodwill when the shipment was rejected. This loss was particularly acute as Fornax's just-in-time inventory system depends on timely and high-quality shipments. Unable to meet its own customers' needs, Fornax had to special-order paper from one of its other suppliers. To make matters worse, this recent rejection was just one in a long series of similar incidents between Great Western Papelco and Fornax. Great Western Papelco has ten paper machines in its system that are nominally capable of producing Fornax paper, but only one, machine C at Pluto, has been "qualified" to supply Fornax, and though qualified, it frequently has trouble meeting the specifications. The operators at the Pluto Mill often take more than a day to change over to Fornax production from the company's own "Great Western High-Speed Reproduction Paper," whose quality specifications, essentially those of the industry at large, are looser than those of the Fornax brand. A paper machine operates continuously so that during these changeovers, a great deal of off-quality paper is made as the operators literally fight the paper machine to bring all 24 key quality characteristics within the customer's specifications. It is true that, although some changeovers have gone remarkably smoothly, on one occasion last year it actually took over 4 days to begin the Fornax run. In contrast, the quality of several other Fornax paper suppliers is apparently high enough that many of their paper machines are qualified, and their quality history is so high that their product is only "skip-lot" inspected. Under such skip-lot plans, Fornax does much less testing, essentially showing confidence in the suppliers' abilities to consistently meet the Fornax needs and specifications. At the Pluto Mill, "Fornax" paper is produced continuously on paper machine C, at a rate of about 20 tons per hour. The linear speed of the machine is about 20 miles per hour. As the paper comes off the dry end of the machine, it is rolled onto huge reels, each of which is about 20 tons in weight, 20 miles long, and 20 feet wide.10 Thus, it takes about 1 hour to produce such a reel of paper. After production, reels are rewound and cut, across the paper machine direction, into 10-mile-long "sets" (still 20 feet wide) and then rewound again and slit four times, again along the paper machine direction, into 5-foot-wide rolls from which the paper is finally cut into 8½ × 11 inch sheets. This last operation is done on a machine called the "WilSheeter," which handles 6 rolls at a time. At the back end of the sheeter, the paper is packaged into 500-sheet reams, labeled, and then packed for Fornax in 20-ream cartons. The run in question consisted of 132 reels, or 2640 tons of paper. Many of the more than 20 quality characteristics measured on each reel of paper are of concern to Fornax and have caused the Pluto Mill problems in the past. The key issues lately, however, appear to have been moisture content, smoothness, and curl. Moisture content was the chief complaint on the last Fornax run. It affects both the printability of the paper and the speed and ease with which it goes through the complex Fornax machines. These stateof-the-art reproduction devices take ream paper in at one end and produce completely bound reports at the other. Their high speed, high quality, and versatility are important competitive edges in Fornax Company's own struggles to regain share in the market it once dominated. The high quality and low cost of the duplication machines offered by several competitors had nearly knocked Fornax out of the market. Part of the Fornax corporate strategy to regain market share is an intensive total quality management effort incorporating a thorough vendor qualification program. Fornax engineers have often visited the Pluto Mill, sometimes during Fornax runs, and it was they who had first introduced some key statistical and quality improvement ideas to Pluto personnel. Although Fornax has provided help and encouragement to Pluto for some time, it appears that their patience is running very thin. Fornax procurement personnel rather bluntly reminded Great Western Papelco's vice president, George Philliston, of their intention to reduce the number of suppliers by half over a 3-year period. By the way, as Great Western well knows, Fornax is able to market reproduction paper wrapped under its own Fornax brand name at a premium price that is well above that of Great Western's own products. Sales of this product make a handsome contribution to Fornax's profitability. Ironically, due to Pluto's difficulty in changing over to Fornax product and its frequent necessity to cull out "off-spec" paper made during the run, it was suspected by some in the mill that they actually lost money on many Fornax runs. On the other hand, it was a help to Great Western sales people to tell other customers for commodity grades that machine C was Fornax qualified. If only they really knew how tenuous that qualification was! Fornax insists on 90% compliance to its 4.0 ± 0.5% moisture specifications (3.5% to 4.5% moisture). The Pluto Mill paper makers claim that achieving compliance to moisture specifications on their own is not too difficult. The problem, they claim, is that moisture is frequently adjusted to bring other characteristics into compliance. Smoothness has been another long-standing concern, and the paper makers say that adjustments in moisture, caliper, and basis weight are continually being made to get smoothness within specifications. ("Caliper" is a measure of paper thickness in thousandths of an inch, whereas "basis weight," recorded in pounds per 500 sheets, is essentially a measure of density.) These quality characteristics are dependent on many factors that are under the nominal control of the paper makers, including the machine speed, a variety of temperatures, pressures and nip clearances, and the distribution and quality of the pulp as it is laid down on the moving wire web at the head of the paper machine. Although there are many parameters under their influence, selecting an optimal control strategy has proven to be far from easy. The effects of many parameters often go in opposite directions. For example, increasing machine speed decreases basis weight while simultaneously increasing moisture. Increasing steam flow after the size press increases both caliper and smoothness, but it decreases moisture. In addition, the physical characteristics of the incoming pulp are crucial and can vary considerably from batch to batch. Paper makers around Great Western frequently say, "Making paper is one-third art, onethird science, and one-third luck with a pinch of black magic. The day nature produces cylindrical, knot-free identical trees is the day paper making becomes pure science!" As part of the organizational architecture of Great Western's Quality Is Everything companywide total quality program, each paper mill has formed a quality improvement lead team. In his role as director of quality assurance, Ed Michaels is the Pluto Mill's lead team leader/facilitator. However, he and his team have, as yet, had very little hands-on experience using the tools of statistical problem solving. No specific quality improvement tasks had yet been selected for attack when mill manager Rich Johnson charged Ed and his team with taking on the Fornax problem. Ed is uncomfortable with starting on such a high-impact, high-visibility problem that has defied the efforts of so many others at the mill for so long, but he has little choice. The Pluto Mill culture is such that when Rich Johnson says "jump" the only question is "how high?" The Pluto quality improvement lead team will meet in the morning to consider the Fornax moisture problem in light of the warnings from the Great Western sales department: Fornax management is so unhappy with recent quality that Great Western is in danger of losing the valuable Fornax account if another lot is rejected. Indeed, it is widely rumored that Fornax continues to give business to Pluto largely to keep them as a back-up against the possibility of supply disruption due to the volatile labor relations in the paper industry. With a large Fornax run scheduled for next week, group vice president Philliston wants to know what actions the Mill is planning to take so that this run meets Fornax specifications.
1. Taking advantage of the statistics training that he just went through, Ed Michaels wants to see what he can learn from the Pluto RRDB (Really Reel Data Base) about past and potential quality problems. Since the last run was rejected because of low moisture, Ed is focusing attention on the data in the file C20_02.xlsx, which gives the moisture measurements from the RRDB for the 132 reels from the rejected run. The moisture measurements contained in this data set are the result of physical tests conducted on a single, 12-inch square sample cut off the end of each reel-at its center "across the reel." It takes approximately half an hour to get the sample to the lab, conduct the moisture test, and report the results back to the machine operator. (Moisture is measured by weighing the sample sheet, baking it for 8 minutes at 100°C and then weighing it again. The weight loss is presumed to be due to evaporation of water. Your assignment is to perform a statistical analysis on this data set that might be useful to Ed and his team in understanding and resolving their problem. The following issues are among those that Ed thinks might be relevant to an understanding of Great Western's problem.
a. Great Western's management has traditionally dealt with quality issues in terms of averages such as moisture averaged over a reel, averaged over a run, or averaged over a month. How well does the average reel moisture over the run conform to the Fornax specifications of 3.5% to 4.5%? How well do the individual reel moistures conform to the Fornax specifications? What should the specification apply to anyway-averages, individuals, samples, or the whole run? What would senior management at Great Western think? At Fornax?
b. Pluto Mill management has reminded the Great Western sales department that traditional industry standards on this grade of paper are 4.0 ± 1.0%. How well do the individual reel moistures conform to these weaker standards? The Pluto Mill manager has complained, "Isn't the problem the inconsistency between Fornax and the rest of the industry?"
c. The next Fornax run is scheduled to be 120 reels. How many of those reels can Ed expect to meet Fornax specifications? How far off could he be in this prediction? Could he give the Great Western sales department a range on the number of "off-spec" reels that might be produced? If the mill were to run an extra number of reels and cull out the off-spec paper, how many reels would they have to run to be certain of getting 120 good ones?
d. Dr. D. Vader of the Fornax quality assurance department devised the sampling inspection plan that was used by Fornax when it rejected the recent run. Vader's starting point was the frequently used MIL-STD 105D plan (which he modified because Pluto ships in FIFO order and Fornax and Pluto are collaborating on a "just in time" inventory program). Thus, the testing at Fornax is done up front and just in time to decide whether the rest of the shipment should be unloaded from the rail cars. In Vader's plan, one carton of paper is selected at random from each of the first 14 reels of paper received from Pluto. Then, 5 reams are selected at random from each carton, and 5 sheets are selected at random from each ream. On the basis of these data (only 14 × 5 × 5 = 350 sheets of paper out of the more than 429 million sheets in the run), Fornax rejected the entire shipment. Mill management is asking Ed how such data could have led to the lot rejection. "Didn't our own tests, made before shipment, show that we met specs?" The Fornax Company has not given Great Western any specifics on their data. All Ed knows is that, according to Vader, 58 of the 350 sampled sheets were below and 1 sheet was above the moisture specs. Aren't these data, Ed mused, inconsistent with the numbers in the RRDB table? What insight can Ed get from his own RRDB records? Is such a small sample enough to assure that Fornax is getting 90% compliance on such a large production run? Are the results-as political pollsters put it-within an allowable margin of error? Might Fornax have made an error in testing? What could explain the differences?
2. As mentioned previously, the next run will be starting on Monday morning-if the paper makers can manage the changeover in a timely fashion. Besides working harder and putting forth "best efforts," what can Pluto Mill do to assure Fornax that there will not be a repetition of last month's disastrous product rejection? Ed Michaels would like to put his statistical training to work to do some root-cause problem solving, but how?
3. Great Western sales and marketing are chagrined at the potential loss of Fornax as a customer. Indeed, they expect that quality standards will be even stricter in the future, and if Great Western is ever to hold on in this product line, or penetrate new markets, real quality improvement will be necessary. It is clear to senior management that the company's quality status has slipped and that Great Western is far from being the quality leader it once was in the industry. Perhaps the Fornax problem is a blessing in disguise. Pluto Mill, together with corporate engineering, has developed a proposal to spend $23 million to upgrade the wet end of the paper machine in an attempt to cure a variety of performance problems, most particularly "curl." Curl also affects how easily paper runs through a duplicator without jamming. It has been a frequent problem with Fornax. Bigabeta, Inc., a leading supplier of paper machine equipment, assures Great Western that the proposed upgrade to machine C will, among other things, ease the Fornax moisture and smoothness problems. Unfortunately, the recent history of such capital projects at Great Western has been disappointing, to say the least. Over the last 10 years, fully 60% of funded projects have failed to live up to their projected impacts. The Great Western board of directors is uneasy about the efficacy of their capital funding process. Something appears to be broken in how these projects are approved or implemented, but what is it?
a. What light do the data in the file C20_02 .xlsx cast on the worthiness of this capital proposal for the moisture problem? Of course, in reality a variety of financial and marketing analyses would be necessary, but in the spirit of the case, focus on the process performance itself.
b. In addition to the types of financial information that usually accompany such capital requests, what other data would you request in evaluating this proposal?
c. Ed himself is no longer as confident as he once was that technology itself will be a cure to the problems on machine C. He wonders, "Is this paper machine potentially capable of meeting the Fornax specifications? Is the problem in the paper machine, or in how the paper machine is operated? Could the problem be (as claimed by the machine operators) in the pulping process upstream? And how does all this relate to the smoothness, abrasion, and curl problems they have had on Fornax paper over the past several months?"
4. From the Fornax view, what do you think about the Fornax specifications and testing? Do you have any suggestions for improvement?