Feeding The Gaming Addiction With In-Memory Analytics

How has real-time analytics made a difference to the performance of mobile games?


Mobile gaming is big business. Candy Crush Saga players spent a whopping $1.33bn on the game in 2014, and over $1bn in the second half of 2013 alone. However, numbers have steadily been falling as other companies realize the profit there is to be made and enter the market in droves, giving customers a vast number of options as they tire of Candy Crush. In the face of such intense competition, the game’s manufacturer King Digital Entertainment has turned to in-memory analytics to gain an edge over its rivals.

King Digital Entertainment has in excess of 195 games on its books, which are available in over 200 countries. Such a massive amount of gameplay generates a massive amount of data - more than one petabyte a year. And as the popularity of their games grow, so too does the amount of data available to analyze.

Data scientists have analyzed this data to leverage actionable insights, which they can use to understand what is causing users to abandon a game in real-time, identify other players at risk of leaving and find the most valued customers. Creatives at King can then improve the gaming experience for customers and come up with player retention strategies to try and prevent such players quitting. Analytics in a mobile setting also helps to pinpoint players who represent the highest value in terms of propensity to purchase virtual goods, to optimize in-game virtual goods sales.

King’s data scientists use a combination of open-source Hadoop and an in-memory MPP database, which enables the data to be analyzed in real time so they can stay ahead of customer’s needs. Vincent Darley, VP of Data Analytics and BI for King Digital Entertainment, explained some of the metrics they use to do this while speaking at Computing's Big Data & Analytics Summit. He told the audience: ‘One measure is second day retention; of the people who install the game today, how many come back tomorrow and play. And another one is 14 day retention; people who've installed the game today, how many come back two weeks later and are still playing.’ Darley and his team made tweaks to the game to gauge what impact the changes had on these metrics. They found that second day retention fell, but by letting the test run they discovered that 14 day retention was higher, so in the long term, customer experience was better.

Andy Done, data platform lead, notes that this was particularly useful when they discovered many players were getting stuck on level 65. They subsequently made the level slightly easier, while maintaining the necessary amount of challenge to keep the game playable.

King is not the only developer using analytics. Zynga, the social video game developer that has Farmville in its stable, is also using them in a big way. Analytics have a number of other benefits that aren’t necessarily in-game, particularly for marketing. Analytics can be used to optimize ad-generated interaction, and a game developer can identify viral users that spread the player base. They can then reach out on social media sites with offers of rewards to make sure certain desirable behaviors are perpetuated.


Read next:

Working At The Boundaries Of Aesthetics And Inference