In today's ever-changing business environment, it is crucial for businesses to market their products with the needs of users in mind to increase profitability. With consumers increasingly receiving services through digital platforms, analyzing potential users of products through big data has attracted extensive attention from researchers. The obvious differences in characteristics between users have greatly increased the difficulty of predicting potential users of a product. Fortunately, machine learning-based data analysis methods offer solutions to this complex task. However, there are obvious differences in usage scenarios and performance between different algorithms, which brings inconvenience to practical applications. This study delves into machine learning-based methods for product prospecting, including k -nearest neighbours (KNN), support vector machine (SVM), decision tree, XGBoost and random forest. Meanwhile, the performance of these algorithms is compared with data to analyse their advantages and disadvantages. Finally, the full paper is summarized and future research directions are envisaged.