Teaching Data Mining: The New “Required Competency” for Marketing ProfessionalsIndustry trendsWhat is Data Mining?Slide 4Slide 5Data mining isData mining is notSize and Demand for DM SoftwareWhy data mining?Types of data mining applicationsTypes of data mining applicationsSlide 12Operational CRM isn’t enoughSlide 14Slide 15Slide 16Slide 17Why data mining in marketing?CRM applications in marketingSlide 20Slide 21Slide 22Slide 23Case studiesClustering techniquesSlide 26Clustering in ClementineAssociation algorithmsSlide 29Sequence associationSample of sequence association outputPrediction & classificationSlide 33Slide 34What data mining has done for…Slide 36Slide 37Teaching Data Mining: The New “Required Competency” for Marketing ProfessionalsToday’s Presenters:Tom NugentKenneth Elliott, Ph.D.Industry trends•Explosive data and information growth•“Predict or perish!”•Industry has higher expectations of new graduates•Soft economy means the most competitive job market in yearsWhat is Data Mining?Discovering meaningful patterns in your dataAs the data grows… What is Data Mining?The relationships become more complicatedWhat is Data Mining?Data mining discovers meaningful patterns in your complex dataData mining is•A user-centric, interactive process which leverages analysis technologies and computing power“Computers and algorithms don’t mine data; people do!”Data mining is not•Blind application of analysis/modeling algorithms•Brute-force crunching of bulk dataSize and Demand for DM SoftwareWorldwide Data Mining Market ($M)02004006008001000120014001600180020002001 2006Source: IDC. 2001Why data mining?•Standard Life secured 50 million in mortgage revenue•Verizon Wireless retained 33% of targeted customers, reduced direct mail budget by 60% and increased usage and revenue•Softmap achieved a 300% year-on-year rise in website profits the first month they deployed models for personalizationTypes of data mining applications •CRM: analytic applications designed to measure and optimize customer relationships (e.g. customer profitability, retention, marketing analysis)•Financial/BPM: analytic applications designed to measure and optimize financial performance (e.g. budgeting) and/or to establish and evaluate an enterprise business strategy (e.g. balanced scorecard).•Operations/Production: analytic applications designed to measure and optimize the production and delivery of a business’s products and services (e.g. demand planning, workforce optimization, inventory analysis, healthcare outcomes analysis).Types of data mining applications•Student Relationship Management- change the vocabulary–Student Retention/Acquisition–Enrollment Management–Surveys–Targeted Marketing•Financial Aid Allocation•Web Analysis•Sales/marketing applications in framework of the customer lifecycle–Basis for “analytical CRM”~75% of Data Mining applications are CRM“Fewer than 50 percent50 percent of enterprise wide CRM initiatives will generate payback by 2004.” Gartner Group“Organizations that don’t embrace analytics as a component of their CRM strategies are ultimately going to fail at CRM.” Meta GroupOperational CRM isn’t enough“Data mining is a way to lift CRM projects into a higher level of return on investment.” Meta GroupOperational CRM isn’t enoughWhat analytical CRM doesMore EfficientAcquisitionLonger LastingRelationshipMore FrequentUp/Cross SellTime RevenueLossLess LossProfitMore EfficientAcquisition More ProfitLonger LastingRelationshipMore FrequentUp/Cross SellTime RevenueLossLess LossProfitWhat analytical CRM doesMore EfficientAcquisitionLonger LastingRelationshipEven More ProfitMore FrequentUp/Cross SellTime RevenueLossLess LossProfitWhat analytical CRM doesWhy data mining in marketing?•How often do our best customers buy?•What motivates customers to make multiple purchases?•How can we ensure long-term loyalty?•How do we attract and retain new customers?•How can we personalize and align offers to achieve maximum ROI?CRM applications in marketing•Understanding customers–Quickly uncover the attributes that define customer behaviors–Profile customers to understand their needs and desires–Results in more relevant and targeted customer communications•For example…predict that a 31-year old single male is likely to respond favorably to a discounted travel offer every 6 monthsCRM applications in marketing•Develop targeted offers–Identify propensities to purchase certain products–Maximize campaign results through better targeting –Analyze past results to predict future results•For example…predict that a 22-year old woman who lives in Chicago is very likely to purchase a specific new book releaseCRM applications in marketing•Match specific offers to specific individuals–Fine tune messages by marketing channel–Deliver offers based on customer profile–Results in increased campaign ROI•For example, predict that a 35-year old woman with two children is likely to purchase a new toaster every 2.5 yearsCRM applications in marketing•Execute real-time campaigns–Assign scores based on behavior–Provide an immediate offer based on customer specifics–Results in increased response and long term customer value•For example, offer the money market customer on the phone a good rate on a certificate of deposit, based on their profileCRM applications in marketing•Monitor campaign results–Determine how a campaign is doing–Identify ways to improve response–Maximize results by tweaking campaigns mid-stream•For example, offer current cellular phone customers the same offer as new customers, based on feedbackCase studies•Clustering•Association•Sequence association•Prediction & classification•SPSS customersClustering techniquesClustering techniquesClustering in Clementine•Clustering is used to find natural groupings of cases•The cluster results, shown below, show that certain groups or “segments” have a much higher propensity to respondAssociation algorithms+=+=Association algorithmsSequence association1 Home Page2 e-store3 Check-out PageSample of sequence association outputResults of sequence association indicate which items and in what order have been purchase. We see here that if frozen meal and beer were purchased on the last visit, then frozen meal will be purchased on the next visit with a confidence of 87.1%Prediction & classificationEducationno