Customer Stories

Mediatonic

Games Studio Uses the Cloud for Big Data Insight to Improve Player Retention

Share

Email Facebook Twitter LinkedIn
August 01, 2014

Mediatonic creates and manages games on mobile devices, consoles and PCs. Understanding player behavior is crucial in an industry that has moved to a free-to-play model, with revenues mostly coming from in-game purchases. With help from Elastacloud, Mediatonic implemented Microsoft Azure HDInsight to analyze large volumes of big data, the insight from which is used to supply new content to gamers, improving player retention and monetization.

  • “HDInsight is incredibly powerful for us. It allows us to process and summarize huge volumes of data, replacing the need for a traditional data warehouse. We’ve shifted that workload to HDInsight.”

    Dave Bailey
    Mediatonic
    Co-founder and CEO
  • “We’re already seeing a big difference with Microsoft Azure. We really like the way the solution provides us with more detailed insight than we have ever had before.”

    Dave Bailey
    Mediatonic
    Co-founder and CEO

Business Needs

Mediatonic is an innovative games studio that has written and managed more than 100 games for mobile devices, consoles, social networks, and PCs. These games are played by millions, with recent title Amateur Surgeon 3 reaching number one on the iOS free apps chart in several countries. However, the rise of Free-to-play gaming (where users download games for free, with revenue generated via in-game purchases or advertising) has had major repercussions for Mediatonic and the industry. “With Free-to-play games it is vital to engage, retain, and monetize players—that’s how the industry now works,” says Dave Bailey, Co-founder and CEO at Mediatonic. “This means that games must be refined and improved constantly, which is done by analyzing large volumes of player data and A/B testing. With such competition for players, it is hard to profit from Free-to-play games without this level of analytics.”

This use of big data also meant recruitment issues for Mediatonic, which struggled to find employees with the right skills for MapReduce, the programming model and implementation used for processing and generating large data sets on computer clusters. “MapReduce is a developing area and we really struggled to find the talent to build out and enhance the systems at the rate we required,” says Adam Fletcher, Server-Side Engineering Director at Mediatonic. “We were already using C# and Visual Studio in other areas of the business, so being able to use those same technologies across our analytics infrastructure has allowed us to build that expertise internally.”

To meet its goals of improved player retention and increased monetization, Mediatonic required a powerful and scalable infrastructure capable of analyzing vast data sets. “For detailed decisions to optimize content, our previous solution had limitations,” says Bailey. “We needed something more powerful and had to decide which platform to move forward with.”

Solution

After evaluating its options, Mediatonic turned to Elastacloud, certified Microsoft partner and big data specialist, to recommend and help implement the best platform. “We knew from the start that we’d build and operate the technology ourselves, so the question was really what provider to go with,” says Fletcher. “Microsoft put us in touch with Elastacloud, who came highly regarded as one of Microsoft’s most valued partners and who we had worked with previously on a project to move some of our in-game infrastructure to Microsoft Azure.”

Mediatonic had previously used a data warehouse-grade database and the partnership with Microsoft saw this replaced this with an HDInsight Service on the Microsoft Azure platform. The organization used HDInsight to connect the raw data it was collecting from games with the type of reports it needed to produce. Data from games goes to an application running on Microsoft Azure. This is then pushed into the Azure Service Bus service and HDInsight analyzes the raw data and generates aggregated report data. This improves data analysis and allows Mediatonic to write intelligent applications to understand that data. “HDInsight is incredibly powerful for us,” says Bailey. “It allows us to process and summarize huge volumes of data, replacing the need for a traditional data warehouse. We’ve shifted that workload to HDInsight.”

Mediatonic experienced an incredibly smooth implementation, running the previous system (Amazon Web Services) side-by-side with Microsoft Azure to identify and address bugs without impacting operations. “Elastacloud was excellent in offering us training, but we didn’t need much,” says Fletcher. “We had the in-house skills to evolve the platform in the direction we wanted and the whole process has been stress-free, as we collaborated with Microsoft and Elastacloud throughout to achieve our goals.”

Benefits

The change to Microsoft Azure has allowed Mediatonic to use big data to increase player retention and monetization of games, and using C#, has made it easier to create more intelligent applications. “We’re already seeing a big difference with Microsoft Azure,” says Bailey. “We really like the way the solution provides us with more detailed insight than we have ever had before. The Azure team is passionate about accelerating Game Developers and Microsoft's proactive relationship with Mediatonic has been key to the success of our cloud strategy."

Improved Player Retention and Monetization

During the soft launch of Amateur Surgeon 3, substantial time was spent looking at completion rates of surgeries, aiming to ensure a high percentage of players were progressing. Mediatonic was then able to make adjustments to the difficulty and double day one retention. “We can test different versions of a game on different demographics at the same time,” says Bailey. “We are then able to tweak games in response to real-time data and gain an understanding of each player.” At full launch, Amateur Surgeon 3 had an average of 35 minutes playtime per user per day and an App Store average rating of 90 percent or more across more than 10,000 user reviews.

Intelligent Applications

The move to HDInsight saw Mediatonic fundamentally change how they approach data analysis. “Previously we used some Hadoop, but the majority of our reports were generated by querying huge data sets in a traditional data warehouse,” says Fletcher. “We’ve substituted SQL for C# and MapReduce on HDInsight, allowing us to build more intelligent applications to process the data.”

Agility

Going into the project, one of the key requirements for Mediatonic was agility and flexibility. The company operates many different games, each with their own analytics requirements, which often evolve over time.  “It’s very much a process of discovery. The more we learn about our games, the more questions we have,” says Fletcher. “Switching to a MapReduce-centric approach means that we can constantly add new ways of analyzing the data without being limited by a fixed data warehouse schema.”