如何正确选择供应链软件-麦肯锡分析报告.docxVIP

78Gettingsupply chainsoftware rightKishore Kanakamedala, Glenn Ramsdell, and Vats SrivatsanA study of early adopters of supply-chain-management software showsthat it works best where it is needed mostbut is no use as a bandageover awed processes.ver the past decade, companies have invested heavily in wiringtheir supply chains with software designed to manage informationows among their internal operations, suppliers, and customers. From 1999to 2002, vendors sold more than $15 billion in supply-chain-managementsoftware licenses the rst step in a process. And that gure doesnt includethe cost of the expensive installation and maintenance contracts to come.The software and its developers companies, such as SAP and Oracle, thatsell applications for resource planning, as well as supply chain specialistssuch as Ariba, i2 Technologies, and Manugistics have attracted attentionbecause the software promises to track and predict customer demand and,as a consequence, to adjust the ow of goods more precisely. When supply-chain-management software works, it can help cut inventory levels, improvedelivery schedules, and ensure that supply meets demand all of whichshould make customers more satised. One supply chain executive calls thesoftware oxygen for his business.But the results have been mixed. Some companies (including Dell Computerand Wal-Mart) have harnessed the power of this technology to improve theirsupply chains or, in certain cases, to transform their business models. Othershave run into trouble, some publicly acknowledging that they have spenthundreds of millions of dollars with no impact.OBILLY JONES80T H E M c K I N S E Y Q U A R T E R LY 2 0 0 3 N UM B E R 1Good supply chain management is fundamentalA supply chain, simply stated, comprises the ow of a companys products,information about them, and the money that changes hands between thecompany and its suppliers and customers. When a company manages theprocesses that support these ows well, it can ll orders from customersquickly yet keep its inventory to a bare minimum. When it doesnt, thesupply chain breaks down. Customers place orders but hear that the prod-ucts are out of stock, even when they might be available somewhere in thecompany. Raw materials fail to show up on loading docks, and shipments tocustomers arrive late. Factories run below capacity because customers canceltheir orders. The result can be a quick drop in sales, lost margins, and inven-tory write-offs.Software on its own cant x basic shortcomings in supply chain manage-ment; in fact, it can make things worse. The real benet comes from repair-ing broken business processes, and companies that tackle them beforeinstalling the software can reduce inventory levels and predict demand moreaccurately. This effort alone can increase revenues. Add the software, and theimprovements are accelerated and sustained.But in reality, many companies that have installed supply-chain-managementtools are unhappy with the outcome, so here we focus on helping managersturn around sloppy installations. Companies that gure out how to make thesoftware and improved processes work together to create a more efficientsupply chain will see a better return on their investment.Results differ widelyTo nd out what makes supply-chain-management-software installationssucceed, we studied the records of 63 high-tech companies in the Fortune1000 from 1995 to 2001, a six-year period that coincided with the rst bigwave of investment. Our sample included companies that make hardware,networking equipment, semiconductors, and electronics, but not software.Although we concentrated on this group because companies in it were earlyadopters of such applications, we believe that the results of the study applyequally outside the high-tech sector. Additional examples have been drawnfrom other industries.As a metric, we used inventory turns a reliable indicator of the health of asupply chain, since they directly affect returns on invested capital, workingcapital, and the cost of goods sold (through obsolescence and inventory-carrying costs). Supply chain executives face constant pressure to speed upinventory turns. As one told us, “I live or die based on inventory levels. Theproduct portfolio is what it is, and my job is to drive quicker turns.”G E T T I N G S U P P LY C H A I N S O F T W A R E R I G H T81We found that, on average, inventory turned faster for companies thatinvested in supply chain software than for those that didnt. The top third ofcompanies that did install it performed, on average, 100 percent better thanthe lagging third. More surprising, though, was another nding: companiesthat hadnt invested in the software also outperformed the lagging third, andsome even improvedtheir turn rate (exhibit).EXHIBITSupply chain software is not a silver bulletOne electronics manu-Change in inventory turns,1 percentfacturer with severeinventory problems too much stock on300250200Year ofinvestment inHigh-performingadopters (top 1/3)some items, not enoughon others had the150100software2Average-performingadoptersforesight to begin byredesigning its manu-facturing process, in500501 0 1 2 3 4NonadoptersLow-performingadopters (bottom1/3)order to get the mostYearout of its plannedinvestment in supplychain software.122 invested in supply-chain-management (SCM) software at various times during period; inventory-turns analysis begins in 1995 for nonadopters.2Products with stabledemand patterns were produced at steady levels that evenly replenished theinventory; others were produced at levels determined by managers who hadproved their ability to predict demand. Still others, for which demand hadproved difficult to forecast, were made to order. The three production sys-tems were put in place before a cent was spent on new software.In fact, we found that the greatest improvements in supply chain manage-ment occurred when companies both improved their processes and correctlyinstalled and used the software.Making it workThe difference between a successful installation and an unsuccessful one, aswe have stated, is a supply chain that is already in good shape or, at the veryleast, upgraded as the software is installed. In addition, companies must beready: if management, the IT department, and users dont recognize anurgent need to improve the supply chain, they are unlikely ever to get themost out of the software.Such an urgent need to improve supply chain management can arise whenproduct margins shrink, demand drops, customers ask vendors to manageinventory for them, or new products are introduced. Consider the PC assem-bly business, which expanded by about 10 percent a year during the 1990sFor 62 high-tech companies in Fortune 1000 (exhibit excludes 1 outlier) over the period 19952001;For those who adopted SCM software.82T H E M c K I N S E Y Q U A R T E R LY 2 0 0 3 N UM B E R 1but lost margins because of price wars and increased commoditization. Themarket dynamics created a burning platform for these companies, forcingthem to pursue new operating efficiencies. Other companies face the chal-lenge of re-creating that sense of urgency, which seems essential to ensurethat everyone accepts the need for change.A 100 to 150 percent improvement in inventory turnover within two yearsof implementation was common among companies in our sample that notonly installed supply chain software but also looked hard at the way theygathered and used demand forecasts and set inventory levels and at what cus-tomers regarded as important such as on-time delivery or the availabilityof products. If this analysis wasnt carried out before the software wasinstalled, weaknesses in the process tended to be magnied, because whenbad information is owing, software only helps it ow faster. Fortunately forcompanies that dont carry out such an analysis, enterprise software is oftencustomized as it is installed. Companies looking to turn around underper-forming installations can thus move cautiously to adjust individual parame-ters as particular processes are xed.The performance of software, however, depends not just on its featuresand quality but also on the way people use it. With this in mind, we haveidentied the common mistakes that companies make and suggested strate-gies for tackling them. The result is four guiding principles: x only impor-tant broken processes, promise only what you can deliver, improve training,and make people accountable. Since these are the areas that companies gotwrong, xing them improves the softwares performance and delivers abetter return on investment.Fix only the most critical broken processesSome companies undertake massive process redesign programs when theirsupply chains fail to deliver. On the face of it, this must be better than notxing processes at all, but we found that these programs are rarely worththe effort. Instead, companies should x the processes that have the mostimmediate economic impact: forecasting demand, setting inventory levels,avoiding stock-outs (that is, running out of inventory), and ensuring thatdelivery schedules are met. They can then use the software to implementthose improved processes.A PC assembler regarded as its most important broken process the interac-tion between the sales force which made unrealistic forecasts based onquotas rather than realistic ones based on demand and the forecast ana-lysts. The company decided to force the two groups to produce a joint fore-cast. Meetings, tense by design, between the salespeople and the analystsG E T T I N G S U P P LY C H A I N S O F T W A R E R I G H T83revealed the basis for the estimates of the sales force. Once the numbers hadbeen agreed upon, they were fed to the supply-chain-management system fordirect input into factory systems regulating the number of PCs that wouldbe produced. This approach made it possible for the company to cut downdrastically on the manufacture of products that were no longer popular andto reduce stock-outs.By contrast, a computer-networking company implemented supply-chain-management forecasting software while largely ignoring the most importantbroken process: a sales incentive program that encouraged overproduction.The software accurately read the overproduction, but its warnings wereoverlooked because sales were hot a customer could always be found forthe excess quantities and prot margins were high. So while the softwaregave signals that demand was falling, managers and salespeople, blinded byearnings pressure and sales incentives out of step with the companys bestinterests, failed to heed the alert. Raw-materials inventory began topile up and a large write-off ensued. Had the incentivesystem been xed rst, this result might have been avoided.In another case, a consumer packaged-goods companyfaced daunting challenges: 30 percent of its shipments arrivedlate; customers were leaving in droves because it took three days merelyto conrm an order; and inventory that piled up in one factory was out ofstock in another one nearby. The answer to these problems seemed to be amultimillion-dollar investment in supply-chain-management software. Thesize of the investment, though, made the company rethink its supply chainprocesses rst so that it could take full advantage of the software. The com-pany centralized inventory management, began to balance production andinventory among plants, and had suppliers manage their inventory.Only after the company had sorted out these processes did it begin installingsoftware to handle inventory management, production scheduling, ordermanagement, and the planning of manufacturing and distribution. Mindfulof the scale of the change, the company installed the software in sequence soit could capture the benets of each incremental investment it made. As aresult, it improved its on-time delivery rate to 98 percent, increased inven-tory turns threefold, and reduced lead times by a third.Promise only what you can deliverToo many installations win support only after more has been promised tousers than the system can deliver: giving everyone everything they want usu-ally proves too expensive and difficult while slowing down improvementsin high-priority areas. The remedy is to go back and simplify the effort.84T H E M c K I N S E Y Q U A R T E R LY 2 0 0 3 N UM B E R 1One health care company, which built strong support for its supply-chain-management installation through a series of seminars, promised to meet, infull, the wishes of all groups of users. However, the advocates of the soft-ware explained neither how long this would take nor the cost. Six monthslater, those advocates had to admit to the user groups that they would notonly get no more than a small frac-tion of what they wanted but wouldThe greatest value of supply chainalso have to wait much longer thansoftware comes from harnessingexpected for what they would get.its power to improve decisionsSupport for the effort was lost andthe company had to return to thedrawing board. The result was ashift in strategy. The focus moved to managing the launch of new productsand to automating the process of setting prices for products the supplychain problems that were causing the greatest number of headaches and cost-ing most in lost sales and prots. By reducing the scope of the effort, thecompany managed to spend less and deliver more.Improve trainingManagers often focus on automating transactions, something that might beimportant but typically fails to extract the greatest value from supply-chain-management software. The greatest value tends to come from harnessing itsability to improve decisions how much of what product should be shippedto which warehouse, for example by making forecasts more precise. Sincedecision support affects inventory turns and customer-service levels (avoid-ing stock-outs, for example), improving it has more impact than, say, migrat-ing an order process from telephones or fax machines to computer systems,a change that at best reduces head counts.To tap this potential, more training is needed than companies usually antici-pate. Users, even conrmed “nontechies,” must be intimate with the system;they should be taught, for example, how to analyze its results and to performreality checks on its recommendations. One leading semiconductor companythat uses supply-chain-management software for global planning requiresusers to attend a training program covering basics such as the user interfaceand the production of reports. The program also covers more advancedtopics for instance, why the software suggests that the company purchasea certain quantity of raw materials or ship nished goods to this or thatlocation, as well as how to work out the daily demand for chips.Compare this experience with that of an apparel manufacturer that installedsoftware to improve its forecasting and to reduce inventory levels, withouteffect. Senior managers discovered that the companys demand and produc-G E T T I N G S U P P LY C H A I N S O F T W A R E R I G H T85tion planners were feeding “gut feel” forecast numbers and inventory levelsinto the system, ignoring the fact that the software itself could estimateinventory levels. It became clear that many of the planners didnt know thatthe software could predict demand or use different assumptions to recom-mend the production of seasonal items as opposed to everyday ones. Herewas a simple case of a manufacturer that hadnt invested in building the rightskills among planners and thus failed to get the most out of the software.Make everyone accountableWhen large supply-chain-management installations go awry, ngers point inall directions. To avoid such recriminations, top-performing companies maketheir vendors, IT departments, and users jointly accountable while linkingimproved su