【正文】
e and visulization167。 Data mining is feasible, thanks to the puter technology advance167。 CRM Overview167。 Bank CRM features167。 Analytic CRM Process167。 OLAP, Modeling and Data Mining167。 Decision strategies167。 Conclusion OutlineAnalytic CRM – Analysis / Decision Support 167。 All the CRM tools serve the purpose of enhance management decisions167。 Decision strategy utilizes analytic CRM results to optimize business decisionsAnalytic CRM – Analysis / Modeling Case Study167。 Portfolio summaryv Bank of San Francisco Credit Card Program started 6 years agov 950,000 accounts with total balance of $5,342,000 in outstanding receivablev Recently chargeoff rate is high. Annualized chargeoff rate is % v Bank determines to tighten policy based on new model.Analytic CRM – Analysis / Modeling Performance definition167。 Using 12 months performance167。 ExclusionsPast Present10,960 Bads (8% Bad Rate)137,500 Goods4/1998 4/2023 5/2023 5/2023PerformancePeriodPredictivePeriodAnalytic CRM – Analysis / Modeling Univariate Analysis167。 Useful variablesv Account datav Application datav Bureau datav Transactional datav Third party data167。 Top variablesv Utilization in the last 12 monthsv Delinquency in the last 6 monthsv Length of jobv Time of bureau fileAnalytic CRM – Analysis / Modeling Example167。 Rank order all variables 167。 Top 120 variables will go to the next step167。 Use logistic regression to develop model167。 Split population to v Model A: Delinquent customersv Model B: Any delinquent customers167。 Gains chartsAnalytic CRM – Common Mistakes167。 Lack of management support167。 Mismunication167。 Mismatch of business knowledge and technical skill167。 Rush into performance definition167。 Ignore operational limitation167。 Do not understand data167。 Too heavily rely on mathematical model167。 CRM Overview167。 Bank CRM features167。 Analytic CRM Process167。 OLAP, Modeling and Data Mining167。 Decision strategies167。 Conclusion Outline167。 CRM is all about unified customer view167。 Analytic CRM is the brain of all CRM efforts167。 CRM needs to be in line with customer life cycle167。 Data exploration, modeling and data mining are the most important ponents of CRM167。 Avoid ‘Garbagein Garbageout’167。 CRM has be closely link to business practicesConclusion演講完畢,謝謝觀看!