Customer Data Analytics Dr. Goh Khim Yong Infocomm Professional Development Forum 7 th July 2011 Strategic Technology Management Institute 1
Bio and Background Background Business Administration (Ph.D., Marketing) University of Chicago, Booth School of Business Information Systems (M.Sc., B.Sc. 1 st Class) Research National University it of Singapore Marketing and advertising in digital media environments Empirical models of consumer/firm behaviors in the presence of network and social interaction effects Impact of competitive product, pricing and promotional strategies in IT-mediated markets 2
Agenda This short talk aims to introduce Strategic or competitiv ve advantages enabled by customer data analytics (CDA) Skills and experience for analyzing customer data Data analytic tools and database marketing techniques to evaluate e customer relationships Case studies and research findings 3
Strategic / Competitivee Advantages of Customer Data Analytics CDA facilitates the strategic process of Selecting customers a serve firm can most profitably Shaping the interaction ns between a company and the individual customers Coordinating all busine ess processes that help maintain and expand mutually beneficial customer-firm relationships 4
Strategic / Competitivee Advantages of Customer Data Analytics CDA provides insights for terms of competitive differentiation in Providing a comprehensive integrated view of customer base Customizing promotion, information & interactions Customizing offers, products and services Customizing channel outlets, locations Recognizing i needs and wants of defined d customer segments or individual customers Cultivating and developing customer interest, trust and desire Retaining or winning back profitable customers Eliminating unprofitable customers 5
Skills and Experience Needed for Customer Data Analyt tics 6
The New Marketing Landscape The new digital age Recent technology has had a major impact on the ways marketers connect with and bring value to their customers Market research Learning about and tracking customers Create new customize ed products Distribution Communication Video conferencing Online data services 7
Types of Databases (By Information Category) Customer database Prospect database Cluster database Enhancement database 8
Applications of Database Marketing and Data Mining in CRM Uses that directly influence customer relationship Identify and profile the best customers Develop new customers Deliver customized messages that are consistent with product or service e usage Send follow-up messages to customers for post- purchase reinforceme nt Cross-sell products or services Ensure cost-effective communication with customers Improve promotion result by efficient targeting Personalize customer service Engage in stealth communication with customers 9
Building Customer Relationships Changing nature of customer relationships necessitates customer data analytics for Relating with more carefully selected customers uses selective relationship management to target fewer, more profitable customers Relating for the long term uses customer relationship managem ent to retain current customers and build profitable, long-term relationships Relating directly uses direct marketing tools (telephone, mail order, kiosks, Internet) to make direct connections with customers 10
Customer Lifetime Value Multi-period evaluation of a customer s value to the firm NPV of all future profits from a customer over life of a relationship with the firm Recurring Revenues Recurring Costs Principles of CLV Computation Contribution Margin Lifetime of a Customer Discount Rate Lifetime Profit Acquisition Cost Customer Lifetime Value 11
Types of Customer Metrics Traditional Marketing Metrics Market share Sales growth Popular Customer-based Metrics Share of category requirement Size of wallet Share of wallet Transition matrix Primary Customer-based Metrics Acquisition rate Customer Acquisition cost Acquisition Retention rate Defection rate Probability(active) Customer Lifetime duration Activity Win-back rate Strategic Customer-based Metrics Past customer value RFM value Customer lifetime value Customer equity 12
Real-Time and Lead ding Metrics Days since last sale Sales acceleration or deceleration Price sensitivity Number of categories purchased Number of cancelled orders, returns, exchanges Number of enhanced services or entanglement Number of customer complaints Number of customer referrals Communication or interaction patterns Multi-channel shopping ratios compared to norms 13
Social Media Metric cs From: Facebook Marketing: An Hour a Day, by Chris Treadaway and Mari Smith Other social media metrics Volume of consumer-created buzz Sentiment by volume of posts Shift in buzz over time Shift in sentiment before, during, and after campaign Competitive buzz Impact of offline marketing on social marketing buzz Buzz by stage in purchase funnel User-generated content created Growth rate of fans and friends Visits s to store locator pages Rate of virality/pass-along Conversion change due to user ratings, reviews Change in virality rates over time Attendance generated at in-person events Second-degree connections reach Contest entries Influence of consumers reached Numb ber of user-generated submissions i received Customer satisfaction Registrations from third-party social logins 14
Customer Models an nd Analytics Whether to buy/ stay/ respond? Customer retention/ attrition model Logistic regression modeling method Campaign response model Logistic regression modeling method, RFM scoring method What to buy? Product/ brand choice model Discrete choice modeling method Cross-product/ cross-category purchase model Market basket analysis method Market segmentation mod del Cluster analysis method 15
Customer Models an nd Analytics When to buy? Lifetime duration model Survival analysis modeling method Purchase incidence/ duration model Survival analysis modeling method Where to buy? Channel management model Discrete choice modeling method How much to buy? Lifetime value/ profitability model Linear regression modeling method Purchase expenditure model Linear regression modeling Purchase quantity model method Poisson/ Negative Binomial regression modeling method 16
Customer Models an nd Analytics Software packages and tools 17
Case Study 1: Social Media Contents & Consumer Choice A Network Perspective Research questions How do information embedded in social media contents affect consumer purchase choice? How does the impact of social media contents depend on the sources (UGC/MGC) and the means of communications (direct/indirect)? Research context An apparel retailer in Singapore Data source Facebook fan page interaction data between consumers & marketer Customer purchase data fro om the retailer Data analytic methods Text mining, social network analysis Econometric modeling: logistic regression, linear regression 18
Case Study 1: Social Media Contents & Consumer Choice A Network Perspective Variable U_CWF (Favorableness of direct communication from consumer peers) U_DWF (Favorableness of indirect communication from consumer peers) U_CWI 0.154 (Information richness of direct communication from consumer peers) (0.061) 061) U_DWI 0.137 (Information richness of indirect communication from consumer peers) (0.013) M_CWF (Favorableness of direct communication from marketers) M_DWF (Favorableness of indirect communication from marketers) M_CWI (Information richness of direct communication from marketers) M_DWI (Information richness of indirect communication from marketers) Coefficient Std Error 0.274 (0.547) -0.063 (0.043) -0.258 (0.116) 0.103 (0.032) 032) -0.005 (0.018) 0.079079 (0.007) Significant Impact? No No Yes positive Yes positive Yes negative Yes positive No Yes positive 19
Case Study 2: Competitivee Impacts of Social Tagging in Online Reviews Research questions How can we quantify the nature of relationships between product demand and outcomes of co onsumer social tagging? What is the nature of demand rivalry in terms of competitive substitution and/or complementarity effects of online consumer reviews e on aggregate dema and for a product? Research context Restaurants and dining consumers in Shanghai, China Data source Online consumer reviews of restaurants Customers restaurant patro onage and expenditure data Data analytic methods Spatial statistical analysis Econometric modeling: linear regression 20
Case Study 2: Competitivee Impacts of Social Tagging in Online Reviews Descriptive analyses Spatial Distribution of Weekly Revenue 21
Case Study 2: Competitivee Impacts of Social Tagging in Online Reviews Descriptive analyses Spatial Distribution of Number of Tagged Recommended Dishes 22
Case Study 2: Competitivee Impacts of Social Tagging in Online Reviews Descriptive analyses Spatial Distribution of Valence of Review Ratings 23
Case Study 2: Competitivee Impacts of Social Tagging in Online Reviews Research findings Volume and valence ratings of reviews for own restaurant are positively related to customer demand d Valence rating of reviews for rival restaurants of the same price category is negati ively related to customer demand Volume of reviews for rival restaurants of the same cuisine type is positively related to customer demand Volume of reviews for rival restaurants within 500m distance and of the same price category is positively related to customer dem mand Valence rating of reviews for rival restaurants within 500m distance and of the same cuisine type is negatively related to customer demand 24
The Analytic Journe ey Starts Numbers Imagination Analysis Logic Left Right Colour Databases Rhythm Sequence Day dreaming As left brain marketing unfolds: Applied science eclipses applied creativity Repeatable processes replace ad-hoc programs Precise analysis repla aces fuzzy measurement Statistical models replace focus groups 25
Contact Information n Have Fun Analysing and Learning About Your Customers! Dr. Goh Khim Yong Email: gohky@ @comp.nus.edu.sg Website: www.comp.nus.edu.sg/~gohky Facebook: www.facebook.com/gohkhimy Phone: 6516 2832 Office: NUS S, COM2 #04-33 26