ThetitleofthethesisIresearchedis“Executionquality:Ananalysisoffulfillmenterrorsataretaildistributioncenter”,whichwaspublishedintheJournalofOperationManagement,oneoftheSSCIjournals.Theanalysiscanbedividedintotwoparts.Inthefirstpart,Ianalyzethestructureofliteraturereview.Inthesecondpart,Igivesometipsaboutthewritingofliteraturereviewbasedonpreviousanalysis.ThefirstpartThestructureofliteraturereview1、previousstudiesabouttheresearchFirstofall,writersreviewedpreviousempiricalstudiesofexecutionqualityinsupplychains.Atthesametime,theyillustratedthedifferencebetweentheirownstudiesandpreviousstudies.Insteadofexaminingshippingdiscrepanciesamongsupplierstoanautomanufactureranddemonstratetheimportanceofelectronicdatainterchange,theystudiedthelinkbetweenaretaileranditssuppliersandexaminedthetypesoffulfillmenterrors.2、theexpansionofpreviousstudiesFromthispart,wecanlearnthatwritersproposedaninventorymodelwhichextendspreviousresearchregardingfulfillmenterrors.Insteadofassumingthattheportionofdefectiveitemsproducedisconstantwithrespecttothelotsizeorthatdefectiveitemsaretheresultofaproductionprocessthatgoesoutofcontrolwithafixedprobabilityperunitproduced,theyfoundthattheincidenceoffulfillmenterrorsdecreaseswithorderquantity.3、theinnovationbasedonpreviousstudiesInthispart,writersexaminedthecharge-backincentivesusedbyretailerstopenalizesuppliersforfulfillmenterrors.Writersdidnotstudycontractingforproductqualityexceptthecostoflapsesinexecutionquality.Inaddition,theyalsoevaluatedwhetherobservedcharge-backsalignsupplychainincentivesbycapturingthecostoffulfillmenterrorsornot.ThesecondpartTips:1、Whenyouwritetheliteraturereview,youshouldbecriticalbutobjective.2、Youshouldintroduceandreviewtheresultsofpreviousresearchesconcisely.(around30words)3、You’dbetterlistyourresearchquestionsorhypotheses.Executionquality:AnanalysisoffulfillmenterrorsataretaildistributioncenterNathanCraiga,,NicoleDeHoratiusb,YanJiangc,DiegoKlabjancaFisherCollegeofBusiness,TheOhioStateUniversity,UnitedStatesbBoothSchoolofBusiness,UniversityofChicago,UnitedStatescDepartmentofIndustrialEngineeringandManagementSciences,NorthwesternUniversity,UnitedStatesAbstractPurchaseordersspecifymanyaspectsofafulfillmentprocess,includingitemquantity,deliverytime,cartonlabeling,barcoding,electronicdatainterchange,retailticketing,andothers.Thesefulfillmenttermsareinstrumentalforhighlyoptimizedretailsupplychainsemployingautomationandtechniquessuchaspack-by-store.Whenfulfillingapurchaseorder,asuppliermaycommitafulfillmenterror,i.e.,thesuppliermayfailtoadheretothetermsspecifiedbytheretailer.Theretailermaythenpenalizethesupplierforthefulfillmenterrorviaachargebackdeduction,whichreducesthesupplier\'srevenue.Wepresentastudyofthefulfillmenterrorsandchargebacksthatoccurinpracticeusingdatacollectedfromamajorretailer\'sdistributioncenter.Whilefulfillmenterrorsinvolvingincorrectproductquantitiesanddeliverytimeshavereceivedthemostattentionintheliterature,wefindthatthemajorityoffulfillmenterrorsinthecontextwestudyinvolvedocumentation,barcoding,andretailticketing.Werefertotheseascorrectablefulfillmenterrors,sincetheyareamendedattheretailer\'sdistributioncenterthroughrework.Wedevelopamodelofinventorymanagementwithcorrectablefulfillmenterrorsandusetheretailer\'sdatatoassessthecostofthesecorrectablefulfillmenterrorstotheretailer\'sinventorysystem.Ourresearchprovidesguidancetomanagersinidentifyingproductsandsuppliersthatimposelargefulfillmenterrorcostsaswellasinsettingappropriatechargebacksforfulfillmenterrors.1.IntroductionRetailersreplenishinventorythroughsupplynetworkscomprisingsuppliers,retaildistributioncenters(DCs),andretailstores.SupplierstypicallyshipproductsorderedbyretailerstoDCs,whichbreakbulkanddistributeitemstoretailstores.InventorymanagementatDCsiscomplicatedandchallenging(deKosterandBalk,2008),notonlybecauseofproductvarietyandlongleadtimes,butalsoduetosuppliershipmentsthatdonotconformtothetermsspecifiedbyretailers’purchaseorders(POs).Thesetermsstipulateitemquantities,deliverytimes,cartonlabeling,electronicdatainterchange,itempackaging,barcoding,retailticketing,andotheraspectsofthefulfillmentprocess.Ifthesupplierfailstoadheretotheseterms,thesuppliercommitsafulfillmenterror.RetailsupplychainsareincreasinglyreliantonfulfillmentthatcomplieswiththetermsspecifiedbyPOs.Supplychaintechniquessuchaspack-by-storeandcross-dockingrequireproductpackaging,retailticketing,andcartonlabelingthatconformtospecifications.Automation,suchasautomatedstorageandretrievalsystemsandtherobotsproducedbyKivaSystems,isunabletofunctionwithoutappropriatebarcodingandelectronicdatainterchange.Lapsesinexecutionquality,suchasthefulfillmenterrorsidentifiedabove,impairtheoperationofhighlyoptimizedsupplychainsandunderminefirms’investmentsintechnology.Fulfillmenterrorsalsonegativelyaffectsuppliers.Retailerspenalizefulfillmenterrorsthroughchargebackdeductions,orchargebacks,whichreducesupplierrevenue.Thesepenaltiesrepresentasignificantcostforsuppliers,reducingoverallsupplierrevenueby2–10%(Zieger,2003).Moreover,retailersandsuppliersregularlycontestthecostoffulfillmenterrorsthroughthedisputesandnegotiationsthatsurroundchargebacks(ChainStoreAge,2002).Thefirstgoalofourresearchistobuildonexistingstudiesofexecutionqualitybydescribingthefulfillmenterrorsandchargebacksthatoccurinpracticewithinaretailsupplychain.Arecentsurveyof42retailersfoundthatanaverageof13%ofshipmentsreceivedbytheretailershadinaccurateadvanceshipnotices,orASNs(Retail,2010).Wecollecteddataonpurchaseordersandfulfillmenterrorsfromamajorretailer,Omega.1Thesedataincludeauditreportsthatrecordthetypeoffulfillmenterrorthatoccurred,thechargebackleviedagainstthesupplierfortheerror,andthetimerequiredforOmegatoperformreworkonitemsaffectedbytheerror,ifapplicable.Wefindthat7%ofOmega\'sPOsexperiencedafulfillmenterror.Quantityshortages,ticketerrors,andASNerrorsarethemostcommontypes,accountingfor52%ofallfulfillmenterrors.Moreover,weobservethatfulfillmenterrorswithinretailsupplychainscanbeclassifiedaseithercorrectableornon-correctable,dependingonwhethertheyareamendedthroughreworkbytheretailer.Forexample,quantityshortagesandlateshipmentsarenotcorrectablebytheretailer.Ontheotherhand,manyfulfillmenterrorscanbecorrectedbytheretaileralonethroughon-siterework.Ifthesupplierfailstodocumentashipmentproperly(e.g.,withapackinglist)ortransmitanASN,theretailer\'semployeescanmanuallyinspecttheshipmenttoidentifyitscontents.Similarly,ifthesupplierattachesextraneouspackagingorticketswithanincorrectbarcodetoproducts,employeesoftheretaileramendsucherrors.Correctableerrorsrepresentasubstantialportion—56%—ofthefulfillmenterrorsweobserved.Industryreportsmakesimilarobservations.Threeofthefivemostcommonfulfillmenterrorsidentifiedbyalogisticsproviderarecorrectable:invalidASNs,incorrectUniformCommercialCode(UCC)128labels,andincorrecttickets(Ma,2013).Incontrasttonon-correctablefulfillmenterrors,whichhavebeenstudiedasrandomyields(YanoandLee,1995)andleadtimes(Bagchietal.,1986andEppenandMartin,1988),researchonthecostofcorrectablefulfillmenterrorsislimited.Correctablefulfillmenterrorsimposebothdirectandindirectcostsonretailers.Considerthecaseofaticketerror.Theretailerfacesadirectlaborcostandadecreaseinlaborproductivitysinceemployeesremoveincorrectticketsandaffixtheappropriatetickets.Inaddition,theerrorincreasesleadtimeandleadtimevariability,whichcancausedamagingstockoutsattheretailer\'sstores.Whilecertainaspectsofthecostoffulfillmenterrors,suchaslabor,arestraightforwardtocalculate,manyretailers,includingOmega,donotknowtheoverallcostofcorrectablefulfillmenterrorstotheirinventorysystems(Retail,2010).Thesecondgoalofourresearchistoextendpriorresearchonthecostoffulfillmenterrors,whichfocusesonnon-correctableerrors,byproposingastochastic(Q,R)modelthatincorporatescorrectablefulfillmenterrorsandassociatedrework.WestudythismodelvianumericalexperimentsusingparametersestimatedfromOmega\'sdataon13,500replenishmentstockkeepingunits(SKUs).2WefindthatcorrectablefulfillmenterrorsfortheSKUswestudyimposeasubstantialcostoninventorymanagementatOmega,namely,thecostoftheseerrorsisbetween1%and4%oftheoperatingbudgetoftheDCwestudied.Managershavemultipleapproachesforaddressingfulfillmenterrors.First,sincefulfillmenterrorsmaketheinventorysupplyprocesslessefficientbyimposingunnecessarycostsonretailersandsuppliersalike,managerscanworkwithsupplierstowardtheeliminationoffulfillmenterrors(Wangetal.,2014).Kulpetal.(2007)describethevariousmethodsretailersusetocollaboratewithsupplierstoreducefulfillmenterrors.Second,intheabsenceofcollaboration,retailerscanuseincentives,suchaschargebacks,topassthecostoffulfillmenterrorstosuppliers.Chargebacksshouldreflectthecostofvendornon-compliancetoretailers(Aron,1998).Third,retailerscanadjusttheirinventorypoliciestoaccommodatefulfillmenterrors.Theseapproachesarenotmutuallyexclusive.Understandingthecostoffulfillmenterrorsindetailallowsretailerscollaboratingwithsupplierstocommunicatethiscostandtoprioritizeproductsandvendorsthatwillbenefitmostfromareductioninfulfillmenterrors.Further,understandingthecostoffulfillmenterrorshelpsretailersensurethattheirchargebacksareappropriate.WefindthatOmega\'schargebacksaretoolowinmanycasesandpotentiallytoohighinothers,whichsuggestsmisalignedincentiveswithrespecttoexecutionqualityinOmega\'ssupplychain(NarayananandRaman,2004).Finally,retailersthatelecttoadoptanadjustedinventorypolicyoftenincuraper-SKUcostwhenmodifyinganexistinginventorysystem.Forexample,underthesoftware-as-a-servicebusinessmodel,inventorymanagementsoftwarevendorslikePredictixchargeperSKUtomodifytheirstandardinventorysystem.RetailersmaythuselecttomodifytheirinventorypolicyforSKUsthatwillbenefitmostfromthechange.Insum,ourresearchcharacterizesthetypesandprevalenceoffulfillmenterrorsandassociatedchargebacksthatoccurinpracticeatarepresentativeretailer,identifiescorrectablefulfillmenterrorsinretailsupplychainsthataredistinctfromtraditionalrandomyieldsandleadtimes,andconductsnumericalexperimentsusingempiricaldatatounderstandtheimpactofsuchcorrectablefulfillmenterrorsoninventorysystemcost.Therestofthispaperisorganizedasfollows.InSection2,wereviewrelatedliteratureandpositionourwork.WepresentanexploratorystudyoffulfillmenterrorsandreworktimeinSection3.InSection4,weintroduceaninventorymanagementmodelthatincorporatescorrectablefulfillmenterrors.WepresentanempiricalanalysisoffulfillmenterrorsandreworktimeinSection5.InSection6,wereporttheresultsofnumericalexperimentsconductedontheinventorymodelusingparametersestimatedfromOmega\'sdata.Further,weproposeproceduresforidentifyingSKUswithhighcostsoffulfillmenterrors.InSection7,wediscussthefindingsofourstudyandsuggestdirectionsforfutureresearch.2LiteraturereviewOurresearchbuildsonempiricalresearchonsupplychainexecution.Inarelatedstudy,Srinivasanetal.(1994)examineshippingdiscrepanciesamongsupplierstoanautomanufactureranddemonstratetheimportanceofelectronicdatainterchangeinajust-in-timesetting.Incontrast,westudythelinkbetweenaretaileranditssuppliers,andweexaminethetypesoffulfillmenterrorsthatoccur,thereworknecessarytoaddresserrors,andthecostsandpenaltiesimposedduetoerrors.Relatedissuesinvolvingexecutionqualityinretailingincludemisplacedinventory(CamdereliandSwaminathan,2010),inventoryrecordinaccuracy(Heese,2007andDeHoratiusetal.,2008),andbackroominventory(Gaukleretal.,2007andErogluetal.,2012).Inaddition,ourresearchisrelatedtostudiesoftheimpactofexecutionqualityontherelationshipsbetweenfirmsandtheirsuppliers(Wangetal.,2014).Theinventorymodelweproposeextendspriorresearchregardingfulfillmenterrors.Theerrorsweclassifyasnon-correctablehavebeenstudiedasrandomyields(YanoandLee,1995)andrandomleadtimes(Bagchietal.,1986andEppenandMartin,1988).Whileresearchershavestudiednon-correctablefulfillmenterrorsextensively,correctablefulfillmenterrorshavereceivedlittleattentioninthecontextofretailsupplychains.Correctablefulfillmenterrorsarerelatedtoproductionprocesserrorsthatcanbecorrectedthroughrework,butfulfillmenterrorsdifferfromerrorsintheproductioncontext.Priorresearchgenerallyassumesthattheportionofdefectiveitemsproducedisconstantwithrespecttothelotsize(Petersetal.,1988,ZhangandGerchak,1990,SoandTang,1995,Jamaletal.,2004andErogluandOzdemir,2007)orthatdefectiveitemsaretheresultofaproductionprocessthatgoesoutofcontrolwithafixedprobabilityperunitproduced(Porteus,1986andLee,1992).Incontrast,wefindthattheincidenceoffulfillmenterrorsdecreaseswithorderquantity(seeSection5).Wewouldexpectthisrelationshipif,forexample,suppliersaremorelikelytoinspectlargerorders.Totreattheerrorsweobserveinthefulfillmentcontext,ourmodelgeneralizestherelationshipbetweentheincidenceoferrorsandorderquantity.Moreover,incontrasttopriorresearch,thenumericalexperimentswepresentutilizeparametersderivedfromdatasuppliedbyamajorretailer.Thispaperalsobuildsonresearchthatassumesthatleadtimedependsoncharacteristicsofanorder.Inourmodel,thetimerequiredtoperformreworkonfulfillmenterrorsdrivestherelationshipbetweenorderquantityandleadtime.Reworktimeisafunctionoftheincidenceoffulfillmenterrors,whichinturndependsonorderquantity.MoinzadehandLee(1987)studyarelatedinventorysysteminwhichafirmreordersdefectiveitems,andthenumberofdefectiveitemsdependsonorderquantity.MoinzadehandLee(1989)examineaninventorysystemwherethesupplierdeliversanorderintwoshipments,andthetimebetweenthetwoshipmentsdoesnotdependonorderquantity.Ourmodelalsorelatestomodelsoflotsizedependentleadtime(see,forexample,KimandBenton(1995)andHariga(1999)),whichtreatincreasesinleadtimeduetothetimerequiredtoproducelargerquantities.Finally,weexaminethechargebackincentivesusedbyretailerstopenalizesuppliersforfulfillmenterrors,whichareatypeofsupplychainincentive.Researchinthisareafocusesonselectingpricesandothertransfersthatrealizethesupplychainoptimalproductionquantity(Cachon,2003).Inthiscontext,chargebacksarestudiedasamechanismforcompensatingretailersformarkdownsandunsoldinventory(Tsay,2001andLeeandRhee,2008).Inaddition,priorresearchhasstudiedcontractingforproductquality(ReyniersandTapiero,1995,Baimanetal.,2000,Lim,2001andJinetal.,2014).Thecostoflapsesinproductqualityisduetodefects—inparticular,towarrantiesandscrap.Incontrast,westudythecostoflapsesinexecutionquality,andweevaluatewhetherobservedchargebacksalignsupplychainincentivesbycapturingthecostoffulfillmenterrors.