如何正确选择供应链软件

78verthepastdecade,companieshaveinvestedheavilyinwiringtheirsupplychainswithsoftwaredesignedtomanageinformationflowsamongtheirinternaloperations,suppliers,andcustomers.From1999to2002,vendorssoldmorethan$15billioninsupply-chain-managementsoftwarelicenses—thefirststepinaprocess.Andthatfiguredoesn’tincludethecostoftheexpensiveinstallationandmaintenancecontractstocome.Thesoftwareanditsdevelopers—companies,suchasSAPandOracle,thatsellapplicationsforresourceplanning,aswellassupplychainspecialistssuchasAriba,i2Technologies,andManugistics—haveattractedattentionbecausethesoftwarepromisestotrackandpredictcustomerdemandand,asaconsequence,toadjusttheflowofgoodsmoreprecisely.Whensupply-chain-managementsoftwareworks,itcanhelpcutinventorylevels,improvedeliveryschedules,andensurethatsupplymeetsdemand—allofwhichshouldmakecustomersmoresatisfied.Onesupplychainexecutivecallsthesoftwareoxygenforhisbusiness.Buttheresultshavebeenmixed.Somecompanies(includingDellComputerandWal-Mart)haveharnessedthepowerofthistechnologytoimprovetheirsupplychainsor,incertaincases,totransformtheirbusinessmodels.Othershaverunintotrouble,somepubliclyacknowledgingthattheyhavespenthundredsofmillionsofdollarswithnoimpact.supplychainsoftwarerightKishoreKanakamedala,GlennRamsdell,andVatsSrivatsanAstudyofearlyadoptersofsupply-chain-managementsoftwareshowsthatitworksbestwhereitisneededmost—butisnouseasabandageoverflawedprocesses.OGetting078-085_SCMsoftware12/18/024:30PMPage78BILLYJONES078-085_SCMsoftware12/18/024:30PMPage79GoodsupplychainmanagementisfundamentalAsupplychain,simplystated,comprisestheflowofacompany’sproducts,informationaboutthem,andthemoneythatchangeshandsbetweenthecompanyanditssuppliersandcustomers.Whenacompanymanagestheprocessesthatsupporttheseflowswell,itcanfillordersfromcustomersquicklyyetkeepitsinventorytoabareminimum.Whenitdoesn’t,thesupplychainbreaksdown.Customersplaceordersbuthearthattheprod-uctsareoutofstock,evenwhentheymightbeavailablesomewhereinthecompany.Rawmaterialsfailtoshowuponloadingdocks,andshipmentstocustomersarrivelate.Factoriesrunbelowcapacitybecausecustomerscanceltheirorders.Theresultcanbeaquickdropinsales,lostmargins,andinven-torywrite-offs.Softwareonitsowncan’tfixbasicshortcomingsinsupplychainmanage-ment;infact,itcanmakethingsworse.Therealbenefitcomesfromrepair-ingbrokenbusinessprocesses,andcompaniesthattacklethembeforeinstallingthesoftwarecanreduceinventorylevelsandpredictdemandmoreaccurately.Thiseffortalonecanincreaserevenues.Addthesoftware,andtheimprovementsareacceleratedandsustained.Butinreality,manycompaniesthathaveinstalledsupply-chain-managementtoolsareunhappywiththeoutcome,soherewefocusonhelpingmanagersturnaroundsloppyinstallations.Companiesthatfigureouthowtomakethesoftwareandimprovedprocessesworktogethertocreateamoreefficientsupplychainwillseeabetterreturnontheirinvestment.ResultsdifferwidelyTofindoutwhatmakessupply-chain-management-softwareinstallationssucceed,westudiedtherecordsof63high-techcompaniesintheFortune1000from1995to2001,asix-yearperiodthatcoincidedwiththefirstbigwaveofinvestment.Oursampleincludedcompaniesthatmakehardware,networkingequipment,semiconductors,andelectronics,butnotsoftware.Althoughweconcentratedonthisgroupbecausecompaniesinitwereearlyadoptersofsuchapplications,webelievethattheresultsofthestudyapplyequallyoutsidethehigh-techsector.Additionalexampleshavebeendrawnfromotherindustries.Asametric,weusedinventoryturns—areliableindicatorofthehealthofasupplychain,sincetheydirectlyaffectreturnsoninvestedcapital,workingcapital,andthecostofgoodssold(throughobsolescenceandinventory-carryingcosts).Supplychainexecutivesfaceconstantpressuretospeedupinventoryturns.Asonetoldus,“Iliveordiebasedoninventorylevels.Theproductportfolioiswhatitis,andmyjobistodrivequickerturns.”80THEMcKINSEYQUARTERLY2003NUMBER1078-085_SCMsoftware12/18/024:30PMPage80Wefoundthat,onaverage,inventoryturnedfasterforcompaniesthatinvestedinsupplychainsoftwarethanforthosethatdidn’t.Thetopthirdofcompaniesthatdidinstallitperformed,onaverage,100percentbetterthanthelaggingthird.Moresurprising,though,wasanotherfinding:companiesthathadn’tinvestedinthesoftwarealsooutperformedthelaggingthird,andsomeevenimprovedtheirturnrate(exhibit).Oneelectronicsmanu-facturerwithsevereinventoryproblems—toomuchstockonsomeitems,notenoughonothers—hadtheforesighttobeginbyredesigningitsmanu-facturingprocess,inordertogetthemostoutofitsplannedinvestmentinsupplychainsoftware.Productswithstabledemandpatternswereproducedatsteadylevelsthatevenlyreplenishedtheinventory;otherswereproducedatlevelsdeterminedbymanagerswhohadprovedtheirabilitytopredictdemand.Stillothers,forwhichdemandhadproveddifficulttoforecast,weremadetoorder.Thethreeproductionsys-temswereputinplacebeforeacentwasspentonnewsoftware.Infact,wefoundthatthegreatestimprovementsinsupplychainmanage-mentoccurredwhencompaniesbothimprovedtheirprocessesandcorrectlyinstalledandusedthesoftware.MakingitworkThedifferencebetweenasuccessfulinstallationandanunsuccessfulone,aswehavestated,isasupplychainthatisalreadyingoodshapeor,attheveryleast,upgr