What do you find when you look at billions of dollars of in-app purchases? Certainly more nuance, diversity and variation than than “minnows” and “whales”. New megastudy covering almost 3000 titles shows how we spend money in mobile games, showcasing the incredible diversity in ways that games monetize. While most games earn from a few to tens of dollars per paying player, some games see the top percentiles spending tens of thousands of dollars.
White Paper here – Full Research Report here
Mobile gaming has become a popular form of entertainment globally, with billions of people spending time immersed in virtual worlds. However, the rise of microtransactions in the mobile gaming industry has led to concerns about the potential financial burden placed on heavy users of these games. This concern is not helped by the lack of transparency about in-app purchases in games or what spending looks like in games.
Help is now on the way thanks to a massive-scale analysis of transactions in mobile games, which investigates billions of dollars in spending across thousands of games and almost 70 million paying players.
Cluster analysis revealed four types of in-app purchase spend profiles in games: Uniform: spending is distributed equally across players; Sub-pareto: more unequal but not at top Quasi-pareto: the most populous cluster, where all percentiles of spenders contribute but higher perc. are more monetised; and Hyper-pareto: majority of revenue is generated by a small proprotion of high spending gamers. These tend to be most profitable too. We also found when looking at the top 1% of spenders in each game that there are significant patterns of spend of these across the clusters. Top 1% in Uniform spend $19 in their lifetimes. Sub-Pareto: $138. Quasi-Pareto, $660, and Hyper-Pareto over $1700 on average. Unsurprisingly social casino games bring in the most revenue and have the highest concentration of high spenders; in fact the more a game relies on its top 1% for revenue generation, the more these individuals tend to spend.
We reveals a great diversity in the way that spending in mobile games occurs. While the majority of games earn from a few to tens of dollars from a paying player, the study sheds light on the extremity of player spending in certain games – with a few games seeing top percentiles of their players spending tens of thousands of dollars. In some games, the pattern of spending is highly uneven, and where a large proportion of revenue (approximately 38%) stems from just 1% of spenders of these games.
We also found that specific types of games are typified by higher spending, with simulated gambling products (“social casinos”) at the top. Additionally, there is a small subset of games across all genres, clusters, and age ratings in which the top 1% of gamers are highly financially involved, spending an average of $66,285 each in the 624 days under evaluation in the most extreme case.
The work represents a step-change in trustworthy marketplace analysis, providing the first detailed breakdown of spending in mobile games at the level of individual people across a substantial section of the global mobile games market.
The study’s findings have serious implications for the ongoing debate about video game monetization, which has captured the interest of policy makers and the media worldwide. Previous work across academia and industry has not had the kind of data to back up conclusions as is utilized here. The study is already being used as a benchmark for future developments within the industry. Furthermore, we hope the findings provide much-needed foundational evidence to guide future research on video game monetization and its potential effects on player wellbeing – positive as well as negative – and how this varies across different types of games, game designs and monetization mechanics.
Fig. 1. Raincloud plots showing the distribution of spending amongst the top 1% of spenders within all 2,873 games in the sample, split by cluster, age category, and genre. A logarithmic transformation is applied to the y-axis to allow the visualisation of all spending. Each data point represents a single game: For example, the most extreme adventure game has 1% of its spenders invest more than $10,000 each; the least extreme has them invest less than $10.
White Paper here – Full Research Report here
A Trillion Hours of Play
The SDU Metaverse Lab, University of Southern Denmark and international partners have initiated a strategic collaboration with the world’s premier game engine developer Unity Technologies. The collaboration is focused on research exploring gameplay across millions of games titles and forms the largest research project on games that has ever been attempted.
Digital gaming has become a human activity on par with browsing the Internet, eating dinner or hanging out with friends. However, despite the popularity of video games and the hundreds of billions of hours we put into games yearly, how, why and where we play games is still poorly understood.
Reliable data on global gameplay is hard to come by. This means that valuable insights are lacking which the games industry could use to provide games that are enjoyable experiences to a culturally and geographically diverse audience. Now Unity Technologies, developer of the world’s most widely used game engine with over 1.5 million creators, has begun an unprecedented collaboration with researchers from the SDU Metaverse Lab and partner universities.
This initiative will allow independent researchers to shed light on our gaming lives, using datasets gathered from hundreds of countries and hundreds of millions of players, expanding the scale of game research by several orders of magnitude. The data are completely anonymized and their use by the research team is governed by strict ethical oversight.
Dr David Zendle, lecturer at the University of York, comments: “We are incredibly excited to work with Unity. We know so little about gameplay globally, and prior work has generally been constrained to looking at only a few games or countries at a time. However, if we want to understand gaming in general, we need to study how we play games across vast numbers of titles and hundreds of countries. Thanks to Unity, we can now build an understanding of video game play on a global scale, especially for mobile games.”
Dr Anders Drachen, professor at the University of Southern Denmark, notes: “This kind of research has never been done before in games – the scale is staggering. We are seeing new market and play patterns emerge that we simply did not know existed. This is the first time we have been able to analyse which games are popular where, how they are played and how we engage with them, on a global scale.”
The team, which includes Dr Catherine Flick, De Montfort University and Prof Sebastian Deterding, Imperial College London, has already discovered that there are distinct geographical and cultural patterns in which games that are popular where. Starting with mobile games, analysis of hundreds of billions of hours of playtime data from over 180 countries and territories, reveals how the population of these countries play games is incredibly varied but appears to fit within eight distinct profiles. These ‘cultures of game play’ have never been described before and give the global games community valuable new tools for understanding players both globally and locally.
The project is part of Unity Technologies’ Academic Research Partnership’s Initiative and brings together an international, independent group of experts from across academia, with the world-leading research environment across the SDU Metaverse Lab’s and the partner universities. The collaboration initiates a paradigm shift in the cross-disciplinary domain of games research, enabling scientists to explore game playing in ways previously impossible.
The first reports from the collaboration, covering analysis of more than 250 billion hours of play globally, are already publicly available:
Did people play games more heavily during lockdowns – or did more people play games? A nascent evidence base suggests that individuals played video games for longer during lockdowns – and school closures in specific. This increase in playtime per gamer is linked to the potential for lockdowns to induce gaming disorder.
In this longitudinal study we utilise 2 years’ worth of data, spanning 251 billion hours of gameplay from 184 countries, to investigate whether lockdowns did lead to more playtime per gamer. We find that whilst school closures are associated with increases in both total playtime within a country and the number of unique individuals playing games, no lockdown policy can explain substantial increases in playtime per gamer: our telemetry data fails to corroborate public health concerns regarding heavy playtime during lockdowns.
Transnational Patterns in Mobile Playtime
Do different nations play games differently? Previous academic attempts to map international differences in gaming have tended to focus on small groups of nations and self-reports of playtime. This has led to an evidence-base in which the existence of specific national gaming cultures are not well-supported by evidence. In this paper, we conduct a large-scale cluster analysis of playtime across 214 separate countries. Contrary to our expectations, we find that many East Asian countries are not highly different from high-GDP North European countries across several measures of play. Instead, a range of previously unstudied and highly differentiated transnational gaming cultures emerge from the data, ranging from Pacific Island states where gaming is infrequent but heavy; to clusters of small, wealthy east Asian territories (Singapore, South Korea, Macao SAR, Hong Kong SAR) with high standards of living and high levels of daily playtime per capita. This analysis highlights the diversity of gaming cultures across the world, and opens the door to future cross-cultural research in gaming.
Game Analytics Resources 