Nazara is a leading diversified gaming & sports media platform with a presence in India, Africa, and North America. The company has offerings across the interactive gaming, eSports, ad-tech, and gamified early learning ecosystems, including World Cricket Championship (WCC) and CarromClash in mobile games, Kiddopia in gamified early learning, Nodwin, PublishMe, and Sportskeeda in eSports and eSports media, OpenPlay, Halaplay, and Qunami in skill-based, fantasy and trivia games, and Datawrkz in digital ad tech.
The customer realized that to stay competitive, they needed a solution to understand their player's behavior and interactions with the platform, use game telemetry to enhance the player experience and make decisions based on streaming real-time data.
Nazara experienced unprecedented growth and demand from the mobile games segment. They manage about 10 million hourly active users with 600k subscription changes daily while managing traffic data from 75 different telecom operators across Singapore, India, Indonesia, Malaysia, Sri Lanka, Bangladesh, South Africa, and other African countries.
The analytics team's key challenge was managing real-time data, providing Nazara business users and players with an enhanced experience.
Nazara also needed observability and deep monitoring into the Apache Kafka cluster to manage platform operations. Since they didn’t want to perform the undifferentiated heavy lifting of infrastructure management and security while pivoting to mobile gaming, the executive team focused on shifting talent to performing higher value-added tasks instead of day-to-day operational management.
To help Nazara analysts identify patterns and trends in their gaming applications, Mactores data science experts collaborated with Nazara business teams on their datasets. The Mactores Applications Strategy team helped Nazara platform teams to add various data points to the telemetry from the games and user interactions. Mactores Application Engineers built a microservices-based event-driven architecture to support real-time streaming of this telemetry data.
This telemetry system consumed Amazon MSK and Amazon Kinesis Data Analytics to ingest and process the telemetry data in real-time.
Real-time telemetry enabled the Nazara business teams to analyze player interactions with the platform and other players with real-time in-game telemetry. To analyze player sentiment in real-time to adjust critical elements of the game, the Mactores Data Science team built machine learning models. The microservices platform gave Nazara players real-time leaderboards and personalized experiences with loyalty programs and incentives.
Nazara achieved a 30% increase in ARPU (Average Revenue Per User) and 80% renewal rates of the player due to these enhancements. By analyzing player activity, Nazara business teams can better understand player preferences and behavior, schedule more relevant content, commission enticing new game releases, and connect more closely with online players improving the overall profitability of the platform.
Mactores built Kafka-based event-driven microservices-based architecture with Amazon MSK, Amazon Kinesis Data Analytics, Amazon S3, Amazon ECS, Amazon App Mesh, and Amazon Sagemaker, which enhanced user experience by 40%. Mactores integrated the Prometheus endpoint from Amazon MSK with Datadog. Implemented Cruise Control to manage Kafka brokers and balance the traffic in real time, reducing operational costs by 25%.
AI/ML models in production built by Mactores Data Science Team predict the socioeconomic class, age, and gender of anonymous or unregistered players from content viewed, giving Nazara platform advertisers confidence in their return on investments in Nazara games. Mactores transformed customers' gaming engines into an event-driven model, containerized and deployed in Amazon ECS with Amazon AppMesh. Integration of Amazon MSK with ECS enabled event-driven gaming decisions by the gaming engine. Integration of Amazon MSK with ECS enabled event-driven gaming decisions by the gaming engine.
Nazara had researched enough to know they needed to conduct a POC. After engaging with Mactores during an Apache Kafka webinar, we collaborated on developing the scope for a POC using Amazon MSK for real-time data, Amazon EMR to deploy the Apache Spark-based ETL Jobs, and Amazon Redshift for end-user queries.
During the POC, Mactores led the Nazara team through an Immersion Day workshop to provide context and insights into the services being deployed as part of the POC. The POC results proved that the Amazon MSK could handle its peak traffic which resulted in using the Mactores proposed design leveraging recommended Architecture for production.