Customers churn prediction and marketing retention strategies. An application of support vector machines based on the AUC parameter-selection technique in B2B e-commerce industry

E-commerce has provided new opportunities for both businesses and consumers to easily share information, find and buy a product, increasing the ease of movement from one company to another as well as to increase the risk of churn. In this study we develop a churn prediction model tailored for B2B e-commerce industry by testing the forecasting capability of a new model, the support vector machine (SVM) based on the AUC parameter-selection technique (SVMauc). The predictive performance of SVMauc is benchmarked to logistic regression, neural network and classic support vector machine. Our study shows that the parameter optimization procedure plays an important role in the predictive performance and the SVMauc points out good generalization performance when applied to noisy, imbalance and nonlinear marketing data outperforming the other methods. Thus, our findings confirm that the data-driven approach to churn prediction and the development of retention strategies outperforms commonly used managerial heuristics in B2B e-commerce industry.