Game Analytics: Interview with field experts
When we first saw the Game Analytics: Maximizing the Value of Player Data book, we were very excited to see such a valuable printed material which gathers exclusive experience of field workers. We had several questions in mind, and asked authors - Magy Seif El-Nasr, Anders Drachen and Alessandro Canossa our questions. They kindly replied and here’s the result of this nice interview.
Game analytics is like an iceberg - what lies beneath is often misunderstood and unknown at its best. What is Game Analytics?
Let’s first get some definitions in place: While the terminology remains somewhat shaky, analytics is fundamentally a process that is grounded in analysis of data to derive results that can inform or shape theories about a phenomenon. It is composed of a large group of theories, methods, processes, architectures and technologies that are used to transform raw data into meaningful information. This information used in the cycle of game production can support decision-making regarding business, project development, design, etc. Analytics is the *process* of discovery and communication of patterns in data towards solving problems and producing results that can drive action, improve performance, support decision making, decide on monetization strategies, support game design etc. - or just because it is fun!
Game analytics is a term we use to denote the specific application of analytics to games. It is important to realize the analytics is about more than just the users, or in the case of game analytics, the players. It looks at games both as products and as projects. As a product, important areas of focus are user experience, behavior and monetization.
How does game analytics help developers drive revenue?
Being data-driven - if done right - means for a company that it has solid ground on which to make decisions (does not guarantee the right decisions are made). This is what can be used to increase revenue in a number of ways, e.g. optimizing retention or increasing internal effectiveness. This is the way analytics can drive revenue in the broadest sense.
Analytics can be a direct driver of innovation as well, but more commonly act in a supporting role - similar to how analytics can support design.
Talking about analytics in general, what does it require to understand and engage customers? [in terms of technology, talent, domain expertise and knowledge]
Our book, Game Analytics: Maximizing the Value of Player Data, discusses parts of this process. But as discussed above, we are just in the beginning of investigating the value of the current methods and the need to establish new methods to look into users’ behaviors accounting for context and users’ individual differences.
Do you think analyzing the end user behavior could be considered a sort of art? Or is something closer to making business and money?
Claiming that analyzing player behavior is an art entails admitting that it is a purely subjective practice and not reproducible or generalizable. Game analytics as a field has been seen as a scientifically driven method that can be used to derive business decisions and can inform design. The process of analyzing the data to derive design decisions, however, is more of a creative process that requires science and art, as if you treat it as a completely objective process, you would overlook the significant human element that is involved in most forms of data mining. Thus, we would say that game analytics as a process informing design is a craft: in between art and science.
What are the daily tasks for a game analyst? What is it like to be a game analyst?
There are many kinds of analysts working with games, or more specifically analysts operating with different goals. For example, some analysts are monetization specialists and focus on in-game economic analysis, while others analyze markets to inform management and marketing. Larry Mellon, one of the most respected people in the field, noted that game analytics tends to focus on the process of developing games, the performance of technical infrastructure, or the players. There is a tendency to equate game analytics with the analysis of player behavior, but this is just one component of analytics for games.
But, if we focus on those analysts who work with player behavior (i.e. also customer behavior), their primary duties (and we are heavily generalizing here!) revolve around establishing data collection infrastructure, acquiring and storing data, pre-processing and otherwise readying data for analysis, analyzing player data, visualizing and reporting results, and – importantly – communicating these results to a variety of stakeholders. The kind of knowledge game analysts can draw on to solve these challenges ranges from statistics, machine learning, psychology, design, visualization, communication, various other Sciences, and beyond.
Predictive analytics is often mentioned when people talk about analytics and games. How can predictive behavioral analysis help a game developer?
If we focus on the specific aspect of game analytics that deal with the people who play games, there are at least two ways we can view them: as players or as customers. The first perspective focuses on investigating how the game is played and trying to inform game design as a result, the second perspective focuses on investigating monetization channels and trying to inform how to design revenue funnels (we are generalizing a lot here).
Predictive analytics is applicable in both cases, with the general purpose of trying to forecast what all, groups of or any one individual player might do in the future. Typical applications of predictive techniques involve predicting when a player might leave a game, or when a non-paying user becomes a paying user. However, using predictive analytics to evaluate design is also common.
What are the difficulties and challenges in today’s game analytics world?
There are multiple challenges facing game analytics as a field. The process of game analytics is composed of various steps each has its own challenges. These steps include: data collection, data analysis, abstraction of data to develop a narrative delivered to different stakeholders.
Challenges to data collection include selection of features to collect, more challenges of sampling, storage, and retrieval. Challenges to data analysis involves the development of algorithms that will establish better understanding of the data given the variable and changing nature of the game itself that constitute the context within which data is collected. While methods have been developed to investigate analysis and collection techniques, less research has been developed on the abstraction and narrativisation step.