{"title":"Predicting customer churn using machine learning: A case study in the software industry","authors":"João Rolim Dias, Nuno Antonio","doi":"10.1057/s41270-023-00269-9","DOIUrl":null,"url":null,"abstract":"<p>Customer churn can be defined as the phenomenon of customers who discontinue their relationship with a company. This problem is transversal to many industries, including the software industry. This study uses Machine Learning to build a predictive model to identify potential churners in a Portuguese software house. Six popular Machine Learning models: Random Forest, AdaBoost, Gradient Boosting Machine, Multilayer Perceptron Classifier, XGBoost, and Logistic Regression, were developed to assess which one would have a better performance. The experimental results show that boosting techniques such as XGBoost present the best predictive performance. The XGBoost model presents a Recall of 0.85 and a ROC AUC of 0.86. Additionally to the model performance, the study of the model's feature importance revealed that some factors, such as the time to solve a support ticket, the type of application, the license age, and the number of incidents, significantly influence customer churn. These insights can help the software industry key drivers of churn and prioritize retention efforts accordingly.</p>","PeriodicalId":43041,"journal":{"name":"Journal of Marketing Analytics","volume":"18 11","pages":""},"PeriodicalIF":4.0000,"publicationDate":"2023-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Marketing Analytics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1057/s41270-023-00269-9","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"BUSINESS","Score":null,"Total":0}
引用次数: 0
Abstract
Customer churn can be defined as the phenomenon of customers who discontinue their relationship with a company. This problem is transversal to many industries, including the software industry. This study uses Machine Learning to build a predictive model to identify potential churners in a Portuguese software house. Six popular Machine Learning models: Random Forest, AdaBoost, Gradient Boosting Machine, Multilayer Perceptron Classifier, XGBoost, and Logistic Regression, were developed to assess which one would have a better performance. The experimental results show that boosting techniques such as XGBoost present the best predictive performance. The XGBoost model presents a Recall of 0.85 and a ROC AUC of 0.86. Additionally to the model performance, the study of the model's feature importance revealed that some factors, such as the time to solve a support ticket, the type of application, the license age, and the number of incidents, significantly influence customer churn. These insights can help the software industry key drivers of churn and prioritize retention efforts accordingly.
期刊介绍：
Data has become the new ore in today’s knowledge economy. However, merely storing and reporting are not enough to thrive in today’s increasingly competitive markets. What is called for is the ability to make sense of all these oceans of data, and to apply those insights to the way companies approach their markets, adjust to changing market conditions, and respond to new competitors.
Marketing analytics lies at the heart of this contemporary wave of data driven decision-making. Companies can no longer survive when they rely on gut instinct to make decisions. Strategic leverage of data is one of the few remaining sources of sustainable competitive advantage. New products can be copied faster than ever before. Staff are becoming less loyal as well as more mobile, and business centers themselves are moving across the globe in a world that is getting flatter and flatter.
The Journal of Marketing Analytics brings together applied research and practice papers in this blossoming field. A unique blend of applied academic research, combined with insights from commercial best practices makes the Journal of Marketing Analytics a perfect companion for academics and practitioners alike. Academics can stay in touch with the latest developments in this field. Marketing analytics professionals can read about the latest trends, and cutting edge academic research in this discipline.
The Journal of Marketing Analytics will feature applied research papers on topics like targeting, segmentation, big data, customer loyalty and lifecycle management, cross-selling, CRM, data quality management, multi-channel marketing, and marketing strategy.
The Journal of Marketing Analytics aims to combine the rigor of carefully controlled scientific research methods with applicability of real world case studies. Our double blind review process ensures that papers are selected on their content and merits alone, selecting the best possible papers in this field.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
