《《精编》CRM in consumer banking》由会员分享，可在线阅读，更多相关《《精编》CRM in consumer banking（44页珍藏版）》请在金锄头文库上搜索。
1、CRMinconsumerbanking Dr YanningZhangDecember 2001 CRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP ModelingandDataMiningDecisionstrategiesConclusion Outline CRM CustomerRelationshipManagement isallaboutunifiedviewofeachcustomer CRMOverview Products Channel Treatments Customer CRMMarketisstillfastgro
2、wing evenwhenoverallmarketisslowingdown ProjectedCRMmarketwillbeover 10Billionin2003 from 3 3Billiononlyin1999 E commence CRMOverview CRMconceptisnotanewone butitbecameahottopicinlast5yearsAcceleratedconsolidationinbankingindustryRecentmajormergersexpandedcustomerbaseMoreproductsavailabletoconsumers
3、TechnologyallowseasiercollectionandaccessofcustomerdataChinamarketfromERPtoCRM CRMOverview CRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP ModelingandDataMiningDecisionstrategiesConclusion Outline BankCRMCharacteristics BankingisextensiveAlmosteveryfamilyhasabankaccountBankdataisdataintensiveBankda
4、taissensitiveandproprietaryBankdataismulti dimensionalBankindustryishighlyregulated PersonalizedtechnologiescantailormessagetoindividualcustomerCollectdataasmuchaspossiblefromvariouschannels CRMStartingpoint unifiedcustomerview Bank TelephoneBanking ATM Kiosk PersonalBanker OnlineBanking In branchTe
5、ller BankCustomerLifeCycle TargetCustomer NewCustomer InitialCustomer HighPotential HighValue LowValue InvoluntaryChargeoff VoluntaryPayoff ProspectPrescreen Underwriting BankCustomerLifeCycle TargetCustomer NewCustomer InitialCustomer HighPotential HighValue LowValue InvoluntaryChargeoff VoluntaryP
6、ayoff ProspectPrescreen Underwriting BureaudataDemographicdataCensusdata ApplicationdataPricingdataAccountactivationdata BehaviordataMonitoringbureaudata Collectiondata DataCollectedandAnalyzedAlongtheLifeCycle CRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP ModelingandDataMiningDecisionstrategiesC
7、onclusion Outline TwotypesofCRMFront endcustomerexperience OperationalCRMBack endcustomerunderstanding AnalyticCRM TwoTypesofCRM AnalyticCRM OperationalCRM Customer AnalyticCRM Brainofcustomermanagement CharacteristicsofAnalyticCRMExtensivedatacapturing processinganddatasourcelinkingIn depthanalysis
8、andmodelingIntegrationwithoperationalCRMsystem AnalyticCRMprocessisaniterativeprocess 73 U S companiesclaimcollectinginformationaboutindividualsDatacollectionisonlythefirststep Importanttoagreeonwhytocollect sincedifferentpartofcompanieshavedifferentmotivationofcollectinginformation AnalyticCRM Data
9、Process Bankdatawarehouse thebasicbuildingblock AnalyticCRM DataProcess DW DataIntegrationDatawarehouseisthecentralplacetohave360degreeviewofcustomersCommonmistakesindatawarehouseLackoflongtermvisionNocommunicationbetweendepartmentsLackofcooperation only37 sharedatawithotherdivisionsNosufficientqual
10、itycontrolConflictdataentrypointtocreateunnecessaryconflictNosufficientdocumentationfordataelements especiallyformissingvalues AnalyticCRM Analysis CompaniesneedtoagreeuponthemeasurementofsuccessBanksuccessmeasurements CampaignresponserateAnnualizedlossesROA ReturnonAssets ROE ReturnonEquity orROI R
11、eturnonInvestment AnalyticCRM Metrics CRMOverviewBankCRMfeaturesAnalyticCRMProcessOLAP ModelingandDataMiningDecisionstrategiesConclusion Outline KeycomponentsofCRManalysis AnalyticCRM In depthAnalysis DataSummaryisthefoundationInmanycases datasummaryis 70 oftheworkCommonissuesConfirmthedataqualityVe
12、rifymissingvaluesCheckvariablecodingContinuousvariablecodingMissingvaluesSpecialvaluesDiscretevariablecodingAvoidmis intepretnumericvaluesBlankmayhavespecialmeaning AnalyticCRM Analysis DataSummary Dataexplorationiscanbeextensive inmanycases thereareover1000variablestoexploreOLAP OnlineAnalyticProce
13、ssing isoftenusedinvariableexplorationCustomerprofilingisamustTherearemanystatisticalmeasurestodealwithunivariateanalysisMean Average VarianceorStandardDeviationMediumKurtosisEtc Inmanycases variabletransformationisnecessary AnalyticCRM Analysis OLAPandVariableExploration Commonvariableshapes Analyt
14、icCRM Analysis VariableExploration Normal S Shape Linear ExponentialDecreasing Convex increasing Skewednormal AnalyticCRM Analysis Modeling Conceptofmodeling mimicthefuturebypastdata Usethedataavailableandfindpatternsfromthepastandthenapplythepatternsinthemodel Past Present Future PastDataBegins Pas
15、tDataEnds PerformancePeriodStarts PerformancePeriodEnds PerformancePeriod PredictivePeriod AnalyticCRM Analysis Modeling ModelingStages PerformanceDefinition DataCollection Sampling UnivariateAnalysis Modeling ModelValidation ModelImplementaion AnalyticCRM Analysis Modeling PerformancedefinitionNeed
16、toconsiderproperexclusionsPerformancemeasuresiswhatoneneedstopredictPerformancedefinitionneedstobecloselytiedinwithbusinessobjectivesConsensusamongallplayersSamplingBankindustryisdatarich however inmanycases thesamplesindifferentperformanceisunbalancesExamples respondersinmarketingmodel badsincreditriskmodelStratifiedsamplingisasolutionforunbalancedsamplesUnivariateAnalysisUseinformationvaluetheorytoexploredataSegmentation populationsplitInmanycasesthearedifferentpatternswithinsubpopulationwhich
