Do Influencers Influence? -- Analyzing Players' Activity in an Online Multiplayer Game
This work addresses the challenge of accurately identifying influencers in games for marketing and engagement, which is incremental as it builds on existing social network analysis but applies it to a new context with specific findings.
The study tackled the problem of identifying influencers in online multiplayer games by analyzing player social networks in Destiny's Crucible mode, finding that traditional metrics for influencers do not necessarily apply and that central players are distinct from influential players in terms of impact on retention.
In social and online media, influencers have traditionally been understood as highly visible individuals. Recent outcomes suggest that people are likely to mimic influencers' behavior, which can be exploited, for instance, in marketing strategies. Also in the Games User Research field, the interest in studying player social networks has emerged due to the heavy reliance on online influencers in marketing campaigns for games, as well as in keeping players engaged. Despite the inherent value of those individuals, it is still difficult to identify influencers, as the definition of influencers is a debated topic. Thus, how can we identify influencers, and are they indeed the individuals impacting others' behavior? In this work, we focus on influence in retention to verify whether central players impacted others' permanence in the game. We identified the central players in the social network built from the competitive player-vs-player (PvP) multiplayer (Crucible) matches in the online shooter Destiny. Then, we computed influence scores for each player evaluating the increase in similarity over time between two connected individuals. In this paper, we were able to show the first indications that the traditional metrics for influencers do not necessarily apply for games. On the contrary, we found that the group of central players was distinct from the group of influential players, defined as the individuals with the highest influence scores. Then, we provide an analysis of the two groups.