E-commerce Gamification Applications User Churn Prediction

In the e-commerce and retail sector, customer retention has become a challenging and competitive field, considering the current competitive environment. Various applications, including gamification, are being implemented to increase customer loyalty and minimize customer losses. Gamification applications, incorporating elements such as rewards and scoring, can enhance customer interaction on e-commerce platforms and promote customer loyalty. From this perspective, the use of these applications and the continuation of customer engagement are crucial for the e-commerce sector. This article includes a binary classification prediction model that predicts whether customers who have used gamification applications on the Hepsiburada platform will continue to use these applications in the future, along with the model’s details. This model can help e-commerce platforms predict potential losses, enabling them to take preemptive actions before churn occurs to prevent the loss. The input features used in model training include demographic information of customers, past shopping behavior on the platform, counts, volumes, recency, and qualitative information about product views, game and campaign interactions. In the training of models, Decision Tree Classifier, Multi-layer Perceptron, LightGBM Classifier, XGBoost Classifier, Random Forest Classifier, K Nearest Neighbor Classifier model algorithms have been used and compared. It has been concluded that the random forest algorithm, which yielded the highest result, is an ideal algorithm to use in this field.