Utilising machine learning and RFM analysis for customer retention in an online grocery delivery startup

Abstract

Retaining customers is one of the key steps towards a financially successful company. Online delivery businesses need to focus especially hard on retaining customers who they have already managed to convert as consumers have more and more competitors to turn to. Despite available tools and methods, recognising a startup’s uniqueness is vital for designing tailored approaches to address customer churn. This thesis is conducted based on data from an early-stage grocery delivery startup and focuses on providing an actionable framework for its management supporting them with retention efforts. Descriptive analysis methods such as Recency, Frequency and Monetary (RFM) analysis and conventional machine learning such as Logistic Regression, Decision Tree, Random Forest and XGBoosting algorithms have been implemented. The RFM analysis showed that the case study company has an almost equal number of customers who are loyal supporters and those who need activation. The best machine learning results were obtained by applying the XGBoost algorithm to predict customer churn. Additionally, the results of this work have implications for the company’s everyday operations by providing a practical and easily interpretable framework for the company’s management to evaluate customer churn going forward as well.