Steam Measurements

It is estimated that nearly 60 percent of Americans play video games, bringing in annual revenue of over $25 billion for PC gaming alone. There is a notable body of academic work focused on characterizing the gamers that make up this market, examining factors such as age, gender, and economic status. That notwithstanding, the sheer size of this demographic has made comprehensive study of gamer behavior a difficult obstacle to tackle.

To better understand gamer behavior, we have conducted a comprehensive measurement study of the Steam gaming network, one of the largest gaming networks in the world. Unlike previous studies that have relied on sampling techniques, our study is the first complete examination of a major gaming network, measuring all 108.7 million user accounts and 384.3 million owned games. We examine gamer behavior across the dimensions of social connectivity, playtime, game ownership, genre affinity, and monetary expenditure.

As a whole, gamer behavior is highly diverse and characterized by heavy-tailed distributions. Most players exhibit modest behaviors in terms of the number of minutes played per day and the amount of money spent on games, though there is a long tail with outliers, such as gamers who maximize achievements or playtime stats, or gamers who collect games they don't play. We find some strong correlations that show that players tend to befriend those who are similar in terms of popularity, playtime, money spent, and games owned. There is a moderate correlation between how much a game is played and how many achievements it offers. We collect a second snapshot of the Steam network and find that our findings are robust across both measurements. We relate these findings to other relevant studies, including gamer stereotypes, game addiction, and social networking.

Heat map and clustering of Steam users and friendships


  • Mark O'Neill, Elham Vaziripour, Justin Wu, and Daniel Zappala. "Condensing Steam: Distilling the Diversity of Gamer Behavior" In Proceedings of the ACM Internet Measurement Conference. Santa Monica, California, USA: ACM Press. 2016.


Download the full dataset.