Membership Analytics include Engagement Metrics, Customer Graph Analytics and Behavioral Predictive Modeling of Attrition. Steven Simpson emphasized on Advanced Analytics Engagement Metric and its importance in building a framework to gather all of the information available so that the ROI helps the lines of business take action. The value is in making data through analytics very actionable and measurable and to create recurring best practices at the credit union. The framework consists of using an Omni-channel approach which is the way a member engages with the credit union is how the credit union wants to engage with them that is through Digital, Traditional or Executive message delivery system with the proper time and approach balance. Advanced Analysis Methodology is about the 4C’s of data collection, cleansing the data, commuting so as to consume followed by taking action and the relevant measures.
On internal data, there is tons of opportunity to recognize the payday lenders, loan application data, ACH data, and Credit Report Data. Engagement Metric is mainly how different age groups use different products in different ways. Segmentation is everything in this. It is significant to segment data for Direct Marketing Message. More specific need-based messages can then be sent to the segments.
Customer Graph Analytics is Humanizing Data which was explained by Brian Ley. It is about deep historical member relationships that tend to be geographically concentrated. AlphaRank helps in Mapping Offline Social Networks. It deciphers how individuals are influencing each other’s purchases and how people are connected. Social Graphs identify human dynamics and high Influencers in particular customer network. The process is to build customer graph data only on historical transaction data, identifying communities and influencer hotspots for spreading behavior. It is regarding Spreading Product and Service Adoption through Customer Graph. In highly meshed communities, communities correlate loosely with geography and there are important cross-linkages between communities. Influencers have a substantial effect on product cross-sell. Cross-selling influencers shifts adoption curve for entire CU. It is a three-step execution to enhance cross-sell using customer graph analytics- Activate −Influence −Convert. It is essential to Predict Charge Offs supplementing with patterns from transaction data as negative behavior spreads. Fraud Analysis can be done using the social graph to inform suspicious purchases and lower false favorable rates.
Khosrow Hassibi discussed behavioral Predictive Modeling of Attrition in Financial Institutions. The pyramid concerning Intelligent Systems for Decisions consists of Data types (Structured, Semi-structured, Unstructured) at its base followed by Data Infrastructure, Business Intelligence, Advanced Analytics and ultimately Decision at the top. Analyzing in predictive attrition is of great importance in B2C organizations. A vast amount of relevant behavioral data is used to model attrition using the agreed upon definition for the special case. Behavioral Attrition models impact on ‘Net Member Growth’ for CU. The ability to predict attrition in advance provides an opportunity to proactively improve it in the portfolio. The model ‘Attrition Score’ is used to segment customers/accounts on the basis of propensity to churn.
Attrition may be voluntary or involuntary. If voluntary attrition can be predicted then action can be taken on it. It involves creating a list and has a reward and cost associated with it so that an operating point can be chosen. The most relevant data to attrition is Customer Signature. Others include online/mobile data, member support call center, product holdings and member information, bill pay, discount brokerage, transactions, demographic data marketing segments and account balances. Attrition Model Performance engages Insights from Attrition Score for product penetration. The other kind is Online Banking Attrition.