King.com is a European powerhouse for mobile and web gaming that receives over half of its traffic from Facebook. As a platform, mobile gaming is highly monetized and looking at outfits like Zynga it’s also highly lucrative (although still a rocky place to live.) The nature of mobile gaming leads companies to looking at who plays their games, how they play their games, and when players are most likely to spend money on virtual items.
ComputerworldUK interviewed Mats-Mats Eirksson, Kings.com director of data warehousing, about the company’s move from Infobright to Cloudera’s Hadoop. As an analytics platform, Hadoop provides not just a processing system for historical data; but also the ability to watch current and real-time activity by users to help mold their experience and change user interaction on the fly to better adapt to them.
“Analytics is one of the things that made king.com the thing that it is today. In the universe that we operate in, gaming online, it is absolutely essential to know as much as possible about the players and optimise everything,” said Eriksson.
“You can get so much information just by looking at the user behaviour, which you can then use to create better games and better monetisation features. At the end of the day we want to make money out of this.”
Last year, I spoke with Amr Awadallah, CTO of Cloudera, about the gaming instinct in humans and how that drives a multitude of different business ventures as well as potentially civilization-building phenomena. Companies who run games have numerous people interacting with their applications across a multitude of levels and each of these interactions can be recorded, watched, and analyzed. For game companies all of this data flowing through their servers contains both useful and totally inconsequential information—shaking that out can be difficult on the fly, but products such as Hadoop give a huge leg-up towards shaking that out.
Making games challenging but fun must be balanced against player interest—a role for analytics
Eriksson from Kings.com describes their business use of Hadoop as watching numerous “events” occurring in games being played by their customers. By employing the use of the highly parallel system for analytics, they’ve been able to add more events that they can record—which is extremely important as the gaming company continues to develops more games.
He mentions the phenomena that people enjoy games because they present a challenge without becoming frustrating. A company with several game products on the market that might have several thousands different challenge levels between them could have frustrations hiding within them that would be difficult to see during testing. However, in open play a product such as Hadoop might start to discover that many players make it to certain levels of the game and begin to either drop out, blow through that section quickly, or otherwise skip content (demonstrating a level of frustration.)
Every game ever produced has to work on this balance, even Modern Warfare—also discussed by Awadallah in how gaming can be affected by analytics—and it becomes an ever-present problem in need of a solution when there’s a great deal of games all working together in a single business model. No doubt Zynga keeps track of millions of data points a day in Farmville and Castleville and uses that to decide where they’ll tweak what to keep their customers interested; and Kings.com has to watch why people are coming to Facebook, or what games they’re playing on mobile and how that experience changes how much they spend—or what advertisements they click on.
Hadoop presents a powerful solution to all of the above-mentioned needs that a gaming company might have: tracking user experience, watching on-the-fly reactions from players to particular games, looking at how players interact with particular elements of the game. In many ways, Hadoop used for gaming allows a more holistic approach to using crowdsourcing (the players themselves) to show where the game may need better balance or careful consideration.