Predicting player churn in the wild

Abstract

Free-to-Play or 'freemium' games represent a fundamental shift in the business models of the game industry, facilitated by the increasing use of online distribution platforms and the introduction of increasingly powerful mobile platforms. The ability of a game development company to analyze and derive insights from behavioral telemetry is crucial to the success of these games which rely on in-game purchases and in-game advertising to generate revenue, and for the company to remain competitive in a global marketplace. The ability to model, understand and predict future player behavior has a crucial value, allowing developers to obtain data-driven insights to inform design, development and marketing strategies. One of the key challenges is modeling and predicting player churn. This paper presents the first cross-game study of churn prediction in Free-to-Play games. Churn in games is discussed and thoroughly defined as a formal problem, aligning with industry standards. Furthermore, a range of features which are generic to games are defined and evaluated for their usefulness in predicting player churn, e.g. playtime, session length and session intervals. Using these behavioral features, combined with the individual retention model for each game in the dataset used, we develop a broadly applicable churn prediction model, which does not rely on game-design specific features. The presented classifiers are applied on a dataset covering five free-to-play games resulting in high accuracy churn prediction

Customer Lifetime Value Prediction in Non-Contractual Freemium Settings: Chasing High-Value Users Using Deep Neural Networks and SMOTE

“…The incorporation of social [4,31] aspects of user and player behavior in LTV prediction is a beneficial avenue for future research. Further, it will be necessary to verify the present results on further products and games to assess their consistency and to enable the development of cross-product predictive systems [19,41]. Last but not least, we deem it highly relevant to explore additional data sources for LTV prediction.…”

“…This is certainly the case for premium and high-value users in freemium settings. Other examples include the prediction of fraudulent behavior [32], detecting purchase decisions [1] or predicting player churn [2,19,33]. Due to the more or less generalizable approaches to finding optimal parameters, conventional supervised machine learning models usually tend to lean towards the majority class especially when we are dealing with highly unbalanced datasets.…”

“…Game monetization, as standard in free-to-play games, is based on IAPs where the player purchases in-game currencies that can then be used to acquire in-game goods. Before building our prediction models, similar to [1,19,22], we define game agnostic and gameplay features for each player that can be grouped under four categories. In category Telemetry we consider primary activity-related metrics such as number of sessions, rounds and days played or the total play and inter-session time within the window of observation.…”

“…E.g. [2,[18][19][20][21] investigate the prediction of player disengagement/churn, [22][23][24] focus on player retention and [1] predict players' purchase decisions in mobile free-to-play games. Player engagement and purchasing are at the core of players' value to a company, and so is their combined outcome: Monetary LTV.…”

Design the churn analysis on games: a review on techniques for churn analysis

“…The ratio of churners over non-churners as a function of the time determines the churn rate (Fabian Hadiji, Rafet Sifat, Anders Drachen, Christian Thurau, Kristian Kersting, Christian Bauckhaget, 2014). " (Gallo, 2014).…”