The Determinants of Video Game Completion Based on User-Generated Data and Game-Specific Attributes: Evidence from Steam Platform

Recent trends in video gaming industry demonstrate that players are overloaded with the games yet feeling excessively satiated before completing even one game they started. Striving to win players’ attention, developers are faced with inability to define project scopes before production and are forced to make important decisions “on the go” regardless of risks for the project, leading to the majority of games being cancelled before completion. The main challenge lies in allocating resources properly and at the very early stages of game development for additional content which after a certain point may lead to diminishing returns with decreasing number of players interacting with the content. Previous research concentrated on factors that drive players game consumption through studying game reviews, but the reasons why and what factors drive or impede game completion had been rarely investigated. Current paper strives to fill the gap by exploiting the power of user-generated data together with game-specific features to uncover the factors contributing to game completion. The regression analysis on 7,390 video games distributed on Steam revealed core attributes of game completion: 1) user-generated data – several topics extracted from reviews with latent Dirichlet allocation, user-defined tags, number of game followers, players and owners; 2) game-specific features – game age, number of achievements and indie genre. Taking these factors into account along with the respective changes they impose to different game development stages, game developers and project executives can make more informed decisions for better resource allocation and definition of project scope.