Predictive Customer Lifetime value modeling: Improving customer engagement and business performance

Abstract

CookUnity, a meal subscription service, has witnessed substantial annual revenue growth over the past three years. However, this growth has primarily been driven by the acquisition of new users to expand the customer base, rather than an evident increase in customers' spending levels. If it weren't for the raised subscription prices, the company's customer lifetime value (CLV) would have remained the same as it was three years ago. Consequently, the company's leadership recognizes the need to adopt a holistic approach to unlock an enhancement in CLV. The objective of this thesis is to develop a comprehensive understanding of CLV, its implications, and how companies leverage it to inform strategic decisions. Throughout the course of this study, our central focus is to deliver a fully functional and efficient machine learning solution to CookUnity. This solution will possess exceptional predictive capabilities, enabling accurate forecasting of each customer's future CLV. By equipping CookUnity with this powerful tool, our aim is to empower the company to strategically leverage CLV for sustained growth. To achieve this objective, we analyze various methodologies and approaches to CLV analysis, evaluating their applicability and effectiveness within the context of CookUnity. We thoroughly explore available data sources that can serve as predictors of CLV, ensuring the incorporation of the most relevant and meaningful variables in our model. Additionally, we assess different research methodologies to identify the top-performing approach and examine its implications for implementation at CookUnity. By implementing data-driven strategies based on our predictive CLV model, CookUnity will be able to optimize order levels and maximize the lifetime value of its customer base. The outcome of this thesis will be a robust ML solution with remarkable prediction accuracy and practical usability within the company. Furthermore, the insights gained from our research will contribute to a broader understanding of CLV in the subscription-based business context, stimulating further exploration and advancement in this field of study.