CRM in consumer banking

CRMinconsumerbankingDr.YanningZhangwww.cnshu.cn海量资料免费下载CRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP,ModelingandDataMiningDecisionstrategiesConclusionOutlineCRM(CustomerRelationshipManagement)isallaboutunifiedviewofeachcustomerCRMOverviewCustomerCRMMarketisstillfastgrowing,evenwhenoverallmarketisslowingdown.ProjectedCRMmarketwillbeover$10Billionin2003,from$3.3Billiononlyin1999.(E-commence)CRMOverview0123456789101999200120022003CRMconceptisnotanewone,butitbecameahottopicinlast5yearsAcceleratedconsolidationinbankingindustryoRecentmajormergersexpandedcustomerbaseoMoreproductsavailabletoconsumersoTechnologyallowseasiercollectionandaccessofcustomerdataChinamarketfromERPtoCRMCRMOverviewCRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP,ModelingandDataMiningDecisionstrategiesConclusionOutlineBankCRMCharacteristicsBankingisextensiveAlmosteveryfamilyhasabankaccountBankdataisdataintensiveBankdataissensitiveandproprietaryBankdataismulti-dimensionalBankindustryishighlyregulatedPersonalizedtechnologiescantailormessagetoindividualcustomerCollectdataasmuchaspossiblefromvariouschannelsCRMStartingpoint-unifiedcustomerviewBankTelephoneBankingATM/KioskPersonalBankerOnlineBankingIn-branchTellerBankCustomerLifeCycleAcquisitionActivationRelationshipManagementTargetCustomerNewCustomerInitialCustomerHighPotentialHighValueLowValueInvoluntaryChargeoffVoluntaryPayoffProspectPrescreenUnderwritingCollection/RecoveryBankCustomerLifeCycleAcquisitionActivationRelationshipManagementTargetCustomerNewCustomerInitialCustomerHighPotentialHighValueLowValueInvoluntaryChargeoffVoluntaryPayoffProspectPrescreenUnderwritingCollection/Recovery•Bureaudata•Demographicdata•Censusdata•Applicationdata•Pricingdata•Accountactivationdata•Behaviordata•Monitoringbureaudata•CollectiondataDataCollectedandAnalyzedAlongtheLifeCycleCRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP,ModelingandDataMiningDecisionstrategiesConclusionOutlineTwotypesofCRMFront-endcustomerexperience–OperationalCRMBack-endcustomerunderstanding–AnalyticCRMTwoTypesofCRMAnalyticCRMOperationalCRMCustomerAnalyticCRM-BrainofcustomermanagementDataMetricsAnalysisImprove-mentCharacteristicsofAnalyticCRMExtensivedatacapturing,processinganddatasourcelinkingIn-depthanalysisandmodelingIntegrationwithoperationalCRMsystemAnalyticCRMprocessisaniterativeprocess73%U.S.companiesclaimcollectinginformationaboutindividualsDatacollectionisonlythefirststep-Importanttoagreeonwhytocollect,sincedifferentpartofcompanieshavedifferentmotivationofcollectinginformationAnalyticCRM-DataProcessBankdatawarehouse-thebasicbuildingblockAnalyticCRM-DataProcessBillingfileMasterfileTransactionfileApplicationfileBureaufileMarketingfileDWDataIntegrationDatawarehouseisthecentralplacetohave360degreeviewofcustomersCommonmistakesindatawarehouseLackoflongtermvisionNocommunicationbetweendepartmentsLackofcooperation,only37%sharedatawithotherdivisionsNosufficientqualitycontrolConflictdataentrypointtocreateunnecessaryconflictNosufficientdocumentationfordataelements,especiallyformissingvaluesAnalyticCRM-AnalysisCompaniesneedtoagreeuponthemeasurementofsuccessBanksuccessmeasurements:CampaignresponserateAnnualizedlossesROA(ReturnonAssets)ROE(ReturnonEquity)orROI(ReturnonInvestment)AnalyticCRM-MetricsCRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP,ModelingandDataMiningDecisionstrategiesConclusionOutlineKeycomponentsofCRManalysisAnalyticCRM-In-depthAnalysisDataSummaryDataExplorationModelingDataMiningDecisionDataSummaryisthefoundationInmanycases,datasummaryis70%oftheworkCommonissuesConfirmthedataqualityVerifymissingvaluesCheckvariablecodingoContinuousvariablecodingMissingvaluesSpecialvaluesoDiscretevariablecodingAvoidmis-intepretnumericvaluesBlankmayhavespecialmeaningAnalyticCRM–Analysis/DataSummaryDataexplorationiscanbeextensive,inmanycases,thereareover1000variablestoexploreOLAP(OnlineAnalyticProcessing)isoftenusedinvariableexplorationCustomerprofilingisamustTherearemanystatisticalmeasurestodealwithunivariateanalysisMean(Average)VarianceorStandardDeviationMediumKurtosisEtc.Inmanycases,variabletransformationisnecessary.AnalyticCRM–Analysis/OLAPandVariableExplorationCommonvariableshapesAnalyticCRM–Analysis/VariableExplorationNormalS-ShapeLinearExponentialDecreasingConvex(increasing)SkewednormalAnalyticCRM–Analysis/ModelingConceptofmodeling–mimicthefuturebypastdata.Usethedataavailableandfindpatternsfromthepastandthenapplythepatternsinthemodel.PastPresentFuturePastDataBeginsPastDataEndsPerformancePeriodStartsPerformancePeriodEndsPerformancePeriodPredictivePeriodAnalyticCRM–Analysis/ModelingModelingStagesPerformanceDefinitionDataCollectionSamplingUnivariateAnalysisModelingModelValidationModelImplementaionAnalyticCRM–Analysis/ModelingPerformancedefinitionNeedtoconsiderproperexclusionsPerformancemeasuresiswhatoneneedstopredictPerformancedefinitionneedstobecloselytiedinwithbusinessobjectivesConsensusamongallplayersSamplingBankindustryisdatarich,however,inmanycas