Case Study: Kakao Games achieves real-time in-game abuse detection and faster time-to-market with Confluent

Kakao Games Uses Confluent to Track In-Game Abuse in Real Time

Kakao Games, a South Korea–based video game publisher, needed to standardize and analyze real-time game log data from many outsourced third‑party studios to detect in‑game abuse, reduce churn, and speed time to market. After evaluating options, Kakao Games selected Confluent and the Confluent Platform as the event streaming foundation to create a shared, agile pipeline for real‑time log processing.

Confluent implemented a pipeline using Confluent Platform with Apache Kafka and ksqlDB for stream processing and anomaly detection, ingesting roughly six terabytes of filtered game log data per week and operating across some 80 databases. Confluent’s solution enabled real‑time tracking and flagging of suspicious activity (cutting multi‑day analyses and 24‑hour queues down to immediate insights), accelerated time to market, and contributed to increases in active users and revenue for Kakao Games.

Kakao Games

Eugene Lee

Director of Infrastructure Division