Retention Analysis and Player Segmentation based on Gameplay Data

The global gaming market has been expanding at a fast pace. However, the online and mobile gaming industry is struggling with the lower retention rate compared to the other domains. Although improving the retention rate will lead to more profit and reducing costs, the extremely high churn rate makes it difficult to apply the mass marketing strategy. In order to solve this problem, player segmentation is one of the approaches to design appropriate marketing campaigns to improve the retention rate. Therefore, the goal of this paper is to improve the retention rate by cluster and retention analysis. To achieve this, new players needs to be segmented. Player segmentation was conducted by using the Agglomerative hierarchical and K-means algorithms and based on the aggregated player variables which were retrieved from play log. The K-means model with Manhattan method is selected as the best clustering model. New players are categorized into three segments by the model; Churners, Constant players, and Initially excited players. Next, retention analysis for the player segments was conducted. Retention rates were calculated on Day1, Day7 and Day28 for the three segments, and characteristics of each segment were extracted. Based on the characteristics, segmental recommendations to improve the retention rate were suggested for the three segments: constant players, churners and initially excited players. The segmental recommendations allow to adequately approach more specific and narrowed players.