In the US alone, people use 3.1 million GB of data every minute, according to cloud-based operating system Domo's report, Data Never Sleeps – a completely incomprehensible figure and a statistic that is steadily increasing, with its estimates suggesting that, by 2020, for every person on earth 1.7MB of data will be created every second.
As citizens of our intensely digital world, almost every action we make now produces vast quantities of data, and even our 'downtime' – most of which is now, rightly or wrongly, hyper-connected – is a large part of this perpetual state of data generation. That means when you are lying on the sofa binge-watching Netflix, browsing YouTube for funny golden retriever videos, aimlessly scrolling through Facebook or being absolutely trounced by an anonymous-but-foul-mouthed teenager while playing your favorite online game, you continue to release personal data into the world.
In fact, gaming is turning into a particular data-heavy art. According to NewZoo, there are 2.3 billion gamers worldwide, and we can only imagine how much data that equates to. As such, gaming companies are increasingly relying on data to constantly improve gameplay and remain competitive. Ahead of his presentation at the Gaming Analytics Summit at DATAx San Francisco, we sat down with Tian Ding, senior data scientist at Blizzard Entertainment, to talk about the world of gaming and the influence data has had and will continue to have over it.
DATAx: How has the data boom changed gaming?
Tian Ding: Data has become more and more important in the gaming industry. It helps assist game developers in making key business and design decisions. For example, game designers have traditionally worked off their instincts and feedback from either players or testers. Nowadays, when it comes to key in-game feature like game balance and future feature design, data plays a bigger and bigger role in their decision-making.
Data analytics and data science can both make a big impact in multiple areas of the game, from day-to-day analytical questions to new feature implementation using advanced mathematics and machine learning. Data also shines in other non-decision-making projects such as game AI. OpenAI's work on Dota2 and DeepMind's work on StarCraft 2 are some examples.
DATAX: How will data shape the industry over the next few years?
TD: Data will continue to play a very important role in the industry. In fact, with the rapid development of big data hardware and software tech, the need for data will become even bigger. Machine learning will also play a bigger part in the full game development and day-to-day operation cycles, in both game design and business components.
Take Netflix or Amazon product recommendations for example. There are already a lot of built-in personalization components. Most game balance problems will be solved with the aid of data, in particular through advanced mathematics and machine learning models. AI techniques will also be used widely – it could help to reduce game development time significantly. Real-time streaming data analytics will be used in more areas such as esports, where real-time decisions need to be made.
DATAx: In such a competitive industry filled with disruption (as Fortnite has proven), how can gaming companies stay ahead of their counterparts?
TD: Gameplay first and commitment to quality. These are the core values in Blizzard. Everything is based on the success of the experiences we provide to our players. The goal is to make games as fun as possible for as many people as can be reached.
Think globally. A lot of times games need to adapt to the cultures and preferences of players around the globe. A game that's popular in the US may not be that appealing to players on other continents without any modifications.
Meanwhile, game companies need to always think of new cool features as well. Without new elements and lack of constantly-evolving gameplay, a game can feel stale to its players and potentially lose engagement.
Listening to community feedback and thinking of the player experience is also super important. At the end of the day, we develop games to entertain our audience, and they are the driving force of a game's success.
DATAx: What tips would you give to industry newcomers on using data effectively to target your audience with success?
TD: First of all, as a person working in analytics, I feel one should play the game and grow passion for it. This doesn't involve data but is probably the most important. Think of it as your own baby. Such passion helps you better understand the game and can help you identify sweet spots and pain points. This will let you develop your game in a more effective way when combined with data.
Nowadays we typically have way more data than what we actually need to solve our problems. The biggest challenge is often to identify which part of the data is important. To do this, you can think about questions like "if I want to make X in the game better, I need to have Y". Your focus will then be "what data do I need to have Y done". By doing this, we will be able to find the signal amid the noise.
Last but not least, it doesn't only take data to make a game great. In fact, good game design should always comprise of three components: Community perception, game designer instincts and data. They form one cohesive entity and none of them should be overlooked.
DATAx: What gaming trends do you anticipate will shape 2019?
TD: With the success of Fortnite and Apex Legends, we definitely expect to see more battle royale games in the market. Open-world games will remain popular in single and multiplayer RPGs as well.
2019 will also be a big year for console games as the new generation of consoles (PS and Xbox) will be released, and I anticipate there being lots of testing and probing for those new platforms. Loot boxes, subscription models and season passes will continue being the top in-game monetization methods.
DATAx: What are you going to be talking about at the Gaming Analytics Summit in San Francisco this May?
TD: I am going to talk about Hearthstone meta-analysis.
Hearthstone is a digital collectible card game by Blizzard. In Hearthstone, players create 30-card decks and play them against one another, always keeping in mind what is currently top of the meta. Meta describes the trends of deck choices currently seen in the game, particularly in ranked mode.
Understanding the current and past meta is crucial in both game design and business decision-making. To achieve this, we built a data pipeline starting from raw gameplay data to analyze the Hearthstone meta. The core part of the pipeline utilizes unsupervised learning method (clustering) to get deck prototypes from millions of deck content variations.
Still not had your fill of gaming content? Read about five 2019 trends in gaming analytics and join us at the Gaming Analytics Summit, the world's leading conference for gaming analytics, taking place as part of DATAx San Francisco on May 14–15, 2019. Register here today!