Machine Learning Algorithms in Target Marketing Analysis

Abstract

Target marketing places a crucial rule in nowadays business. It allows companies to recommend the right products to the right clients. Companies with existing clients’ data are able to gain customer insights with machine learning algorithms to better target the potential clients during a marketing campaign. This study is written to employ machine learning algorithms to predict the success of a target marketing campaign. The existing data of clients from “Trips & Travel.com” company is used to forecast who will purchase a newly introduced travel package with 10 supervised classification algorithms, such as Ensemble Bagging, Neural Network, and Logistic Regression. With the 4888 records of clients and 19 independent variables, these models are trained to predict whether the clients will have a chance to purchase the new released package. This paper compares the results from the 10 classification models with Area Under the Curve (AUC) and Kolmogorov Smirnov (KS) scores, and the results shows that Random Forest outperforms other 9 models. The result of the comparison implies that only targeting the clients who are classified as the targeted segment can increase the result of marketing campaign dramatically and lower the cost of the campaign at the same time.