Apr 24, 2019

How Social Network Analysis and Data Science Can Improve Social Experiences in Games

,

Our communities heavily influence us, but not every individual has the same amount of impact on a community.

I joined Massive more than four years ago as a Biometric Specialist to help our talented team to reinforce their creative decisions with insights gathered from different sources, like player behavior, opinions, and even bodily responses to game events as players are engaging with them.

I have a background in both academia and the game industry, with over 30 different titles shipped. I have also been tutoring students who wanted to pursue a career in User Research, Human Computer interaction and Data Science before I joined Massive.

Through my extensive network – and with the assistance of our Massive data science team (Alessandro Canossa, Sasha Makarovych) –  we set up a collaboration with a group of well-known researchers from University of York (Christoph Sebastian Deterding and Anders Drachen) and Northeastern University (Casper Harteveld) to examine the effects of communities in games and individuals who have the most influence on them.

The results of this study then helped our developers to create an improved social experience in the game.

In order to investigate the communities, we applied one of the most important methods of mapping and measuring community structures: Social Network Analysis. With this toolset and by using exclusively multiplayer features of Tom Clancy’s the Division, we identified individuals who are at the center of their networks.

We then examined whether these individuals have an impact on their fellow community members in terms of playtime and engagement in group-activities. Finally, we compared their social impact with a random sample of players and a sample of top 1% of our players in terms of performance in the game.

The results showed that not only are these individuals behaving differently from top performers and the average player, but their social behaviors are also imitated by their immediate circle.

The results showed that not only are these individuals behaving differently from top performers and the average player, but their social behaviors are also imitated by their immediate circle.

To show this phenomenon, we isolated players that joined an influencer and monitored their behavior for two weeks before and two weeks after the connection was made.

Just two weeks after joining an influencer’s friends, players showed on average a 30 minutes increase in total playtime and 18% increase in their group playtime. These numbers are much lower or even non-existing for players who joined top performers and a random sample.

Just two weeks after joining an influencer’s friends, players showed on average a 30 minutes increase in total playtime and 18% increase in their group playtime.

Through these studies we also discovered that the impact of the influencers is palpable even after a year of joining their circle – both in engagement with the game for longer periods of time as well as playing in groups.

Using Social Network Analysis metrics allowed us to distill a list of socially impactful players with sustainable influence over playtime and time spent in groups, which in turn helped our game designers to create an improved social experience for our players. Our community developers can also leverage these results to more effectively communicate with and analyze player feedback.

Results of these studies will be published by and presented in Computer-Human Interaction Conference (CHI) in Glasgow on May 2019 and has already been awarded an honorable mention by the editorial board.

In order to make the best game experiences possible, Massive is dedicated to utilize new and innovative ways to know players and create “player centered” games.

  • Are you interested in leveraging data driven insights to push Ubisoft games to the next level? We now have an open position for Digital Media Analyst. Apply here!
  • Read more on the details of the study here.

More to read

Discover more articles