Skip to main content

Discovering Deep Insights Into Customer Retention in Colonial First State (CFS)

Project Member(s): Xu, G., Liu, W., Ou, Y.

Funding or Partner Organisation: Colonial Mutual

Start year: 2015

Summary: Colonial First State (CFS) and AAi are exploring partnership opportunities using deep advanced analytics and data- driven evidence-based models to achieve CFS¿s stated business outcomes. Through a series of formal workshops and deep-dive discussions, four potential projects were identified. After a review with CFS, the ¿Retention Analytics¿ project is selected as the front-runner to launch the initiative, as it possesses the right project attributes o Clear target: Retention of $209M in FY15 (i.e. 30th Jun 2015). Target increases substantially from FY16 ($450M) o Measurable metric: An account attrition rate below 9.6% of the number of accounts closed in 12 months o Definable Project Outcomes: The goal of the Retention Analytics Project is to build an engine that generates deep investor insights and self-learning predictive models to provide CFS with actionable knowledge to support CFS¿s stated retention targets and attrition rates. To achieve this, AAi will develop o Pattern Mining and Discriminative Feature Selection Algorithms o Behavior Analysis and Relationship Models o Retention and Predictive Analytics Models The approaches used in the Development Phase of the project include explorative analysis, descriptive analysis, pattern mining, investor churn risk scoring and investor retention predictive analytics which are further outlined in Section 3.3 The key benefits to CFS are o Reusable models that can be further fine-tuned in the future to include new data sets and self-learning engines o Iteratively reach CFS¿s retention target and attrition rate of closed accounts

Keywords: Pattern Mining Behavior Analysis and Relationship Models Customer Retention Predictive Analytics

FOR Codes: Pattern Recognition and Data Mining, Application Tools and System Utilities