Game On: Kongregate Uses Snowflake and AWS to Deliver On-Demand, Elastic Analytics

Kongregate, a leader in the mobile and PC gaming industry, needed a different data warehouse after going mobile. The company wanted to run analytics on its rich data streams, but its current data warehouse was not up to the task. Kongregate began searching for a data management solution that could separate compute from storage and accommodate JSON data, which led them to Snowflake on Amazon Web Services (AWS). The data engineering team now spends much less time on data management and more time developing while delivering the insights needed for gameplay decisions.

Running analytics on the data generated by mobile was too much for Kongregate’s data warehouse. The company was maxing out clusters and resources. Performance was being affected, and costs were at a breaking point. It was time for a cloud data platform that could accommodate new data formats like JSON and higher volume. The company chose Snowflake to deliver the rich analytics it needed to improve customer lifetime value and revenue growth.

Separating compute from storage was the reason Kongregate chose Snowflake. As it dug into features, it discovered that Snowflake could also meet its on-demand, highly elastic analytics needs. The company can now analyze what kind of people are playing games, which networks are bringing the most qualified users, which are bringing the users that are worth the most, and who has the most flexible bidding so they can secure more users and grow the user acquisition pipeline.

“Adding revenue is a big thing in mobile, and we saw that Snowflake could help us with that,” said Max Murphy, Director of Engineering, Data and Operations at Kongregate.

With Snowflake, Kongregate can watch live operations of a game to see where people are turning out. The company has KPIs for time, experience, and purchasers. IT can slice that data to generate new insights and models, which can be used to deliver better experiences or new games. Snowflake also enables the company to track what purchasers are doing using dashboards and visualizations.

Kongregate also improved its data pipeline, which is now used by all its business units. Events are ingested from Amazon Kinesis into an Amazon Simple Storage Service (S3) bucket. From Amazon S3, the data goes to AWS Lambda and process files for ETL. Kongregate uses AWS Elastic Beanstalk for web-based metadata tools. From there, it goes into Snowflake for the company’s analytics pipeline and then from device to chart. “We were immediately able to go full force into building all these larger upgrades to other parts of our data pipeline,” said Murphy.

After deploying Snowflake, Kongregate’s engineering team saved about 30 percent of their time, and the company doubled the number of mobile games it offered. Snowflake enabled Kongregate data engineers to overhaul the company’s data pipeline and also focus on the SDK and event receiver—things the team did not have time for in the past. “Being able to focus on development and not maintenance is huge. So is rapidly solving issues that once were ETL changes,” said Murphy.