Predicting customer satisfaction using machine learning

Abstract

Customer satisfaction plays an important role in the success of e-commerce platforms. CRM provides a framework and tools for companies to effectively manage customer relationships, enhance customer satisfaction, and improve overall performance on e commerce platforms. Our work focuses on analyzing customer satisfaction using machine learning techniques applied to a dataset obtained from the platform "Olist", an online marketplace in Brazil. The dataset consists of 100,000 apps placed between 2016 and 2018 in multiple markets in Brazil. It includes various factors, including order details, product characteristics, payment methods, and customer demographics. Through exploratory data analysis, we reveal patterns and trends in customer satisfaction, using feature selection and pre-processing techniques to identify influencing factors, and compare results to other studies that used the same database with a difference in identifying inputs. Machine learning algorithms such as logistic regression, decision trees, and random forests are used to develop a predictive model. The main findings of our study highlight the impact of payment methods, shipping times, and customer locations on customer satisfaction. We provide practical implications and recommendations for enhancing customer satisfaction, including improving checkout processes, reducing shipping times, and personalizing customer experiences