A Comparative Analysis of Use of Machine Learning Algorithms for Customer Churn Prediction at Supermarkets in Lagos Nigeria

The retail industry needs customer churn prediction to create successful customer retention strategies.The identification of churn customers in Nigerian supermarkets relies on conventional manual survey methods which take up time and generate limited predictive results.The research investigated customer behavior through transaction data analysis to predict churn while identifying the most effective supervised machine learning models for customer churn risk assessment.The study used transaction data from multiple supermarkets based in Lagos, Nigerian.Retrospective data from Plenty Africa, a digital loyalty platform, was used for this comparative analysis.The research employed predictive modeling methods to study customer characteristics and shopping patterns which helped detect patterns that lead to customer loss.The model performance was evaluated through conventional assessment metrics which led to a comparison for determining which approach performed better.The research showed that machine learning methods succeed in predicting customer departure through supermarket data collection which enables businesses to develop effective customer retention strategies.The research also shows how supermarkets can enhance their operational performance by using data-driven churn prediction systems which generate better results than human-based methods in retail operations.