Anders Drachen

Anders Drachen, Ph.D. is a veteran Data Scientist, Game Analytics consultant and Professor at the DC Labs, University of York (UK).

Online games are of particular interest in game data mining because their underlying business model is highly dependent on understanding player behavior, and currently one of the major forces driving the use of and innovation in data mining in game development. In this post, we take a brief look at the historical background behind the current situation and discuss a few business perspectives.

The current push for data mining player behavior in the industry has to a certain extent been driven by the rise of the social online game – or Free-to-Play (F2P) – genre, as well as the widespread popularity of the Massively Multiplayer Online Game (MMOG) genre. MMOGs have an almost two decades long history, reaching back to games like Meridian 59 and EverQuest. They started getting serious attention in the regular press with the realization that these online, persistent worlds contained intricate economies (Castranova, 2001), and with Second Life (arguably a virtual world, not a game) and notably World of Warcraft, that they had become highly popular. With the evolution of Web 2.0 technologies, notably social networking platforms like Facebook, another type of game also increased in popularity: F2P, with early examples on Facebook including Mafia Wars.

MMOGs and F2P games were different from previous game forms in that they catered to very large groups of players who could interact in real time. MMOGs and many F2Ps are also persistent world games – they are always running – which facilitated the emergence of social communities in these games.

In the past few years metrics-driven development has almost become standard in online games development and –management. Acronyms and terms like ARPU, NOSQL and Big Data are becoming commonplace, and it is likely that most publishers and developers in the online games sphere are highly dependent on analytics and reports to keep their businesses running. While many Key Performance Indicators (KPIs) are common, the level of sophistication in the analytics software and processes vary across the industry. Competition, the cross-over of players between different sectors of the games industry, and the evolution in player communities over time, requires online games companies to field efficient data capture and storage, and the ability to generate KPIs and ad-hoc analysis and reporting.

A lot more could be said about the historic background for MMOGs and F2Ps, but the essence of the matter is that these games need data mining because they have to manage and monetize on a community of players. Generalizing, the essential requirement in the MMOG business model is to keep people engaged so they continue to pay subscription fees. The requirement for F2P games is to convince players to spend money on buying in-game resources.

There are a number of ways to handle this kind of challenge, but fundamentally relate to Business Intelligence management. There are a number of similarities between managing and monetizing on player communities and the management of websites, online forums and web-based communities in general. These, similar to online games, have customers coming and going, interacting with the site and/or people via the site, for shorter or longer periods of time.

Web analytics is the field of research and practice dealing with quantitative analysis of user behavior on the Net, and back when MMOG and F2P models were gaining momentum, there was a lot of knowledge available that could be adapted for use in these – and other – types of games, for example with regards to online advertising and customer retention, and the use of techniques like funnel analysis and cohort analysis to understand the cost of acquisition, retention factors, revenue generation, social factors, etc.

The metrics-driven business practice in online games gained strong traction with the rapid growth of game companies like Zynga, BigFish and Wooga, who had adopted a metrics-driven development practice and became highly successful in short period of time, and the growth of the social application market in general (e.g. Facebook, InstaGram, LinkedIn, Google+, Twitter, Picasa …).

Today, business intelligence has emerged as a key aspect of operating a successful online games company. 


Game Analytics is a cloud hosted solution for tracking, analysis and reporting of game metrics. We will improve quality assurance, bug fixing, game design and monetization of games. Drop us a line on if you want to be part of our open beta programme!

Anders Drachen

Anders Drachen, Ph.D. is a veteran Data Scientist, Game Analytics consultant and Professor at the DC Labs, University of York (UK).

Join a community of passionate game developers, who get our newsletter every week!

Sign up for a free surprise