“The faster we can verify and respond to data, the more benefits will ultimately return to the game users,” said Kim Do-kyun, head of the Data Engineering Team at NeoWiz, explaining the necessity for the company’s data infrastructure transformation.
Game companies are increasingly building environments to efficiently manage and utilize the growing volume of game data. This data, generated in real-time during gameplay, includes user play records, item purchases, currency usage, and points of user drop-off, all of which need to be linked to actual business operations.
The speed at which data is analyzed and reflected in operations influences updates, difficulty levels, and reward designs, making data management capabilities increasingly critical for game companies.
In response to the rising demand for data and analysis, NeoWiz has restructured its data environment. Kim Do-kyun and manager Shin Cheol-kyu from the Data Engineering Team recently spoke with Aju Economy about the changes following the infrastructure transition and future utilization plans.
The Data Engineering Team at NeoWiz collects and manages vast amounts of data generated during game services and external marketing, providing foundational data that can be used by practical departments for business planning and operations.
As the volume of data processing increased due to new game releases and major updates, NeoWiz adopted Snowflake’s services last year. After conducting proof of concept (PoC) across five areas and 34 scenarios in the first quarter of 2025, the system is now actively utilized in daily operations.
The most significant change has been in the way data is utilized. Kim noted, “Previously, each department had to request the necessary data, which would then be delivered, followed by additional requests for further analysis. Now, as long as departments are granted access rights and policies, staff can directly use the data they need.”
NeoWiz has also implemented AI-driven analysis methods. Even those not in the data analysis team can now ask for necessary data in natural language and receive results. This has broadened the range of data used in practical applications and diversified its utilization methods. Some data analysis tasks that once took 1-2 weeks have been reduced to about 10 minutes.
Manager Shin emphasized, “To reduce incorrect analyses (hallucinations) by AI, we have put significant effort into defining relationships and standards between data in advance, creating reliable data that can be trusted.”
As the scope of data utilization expands to practical departments, the role of the Data Engineering Team has undergone a major transformation. The team has shifted focus from repetitive data extraction and infrastructure management to more advanced tasks such as cost management and building environments for collecting and analyzing data for new games.
Shin explained, “Now, we are more focused on how to manage and utilize the data we have rather than just preparing it. The nature of our work has changed so much that it could be considered a completely new profession.”
NeoWiz is quickly responding to real-time data analysis by applying the new analytical system to web board games like ‘Pmang New Matgo’ and live service games such as ‘Brown Dust 2’ and ‘Cat and Soup.’ For console titles like ‘Lie of P’ and upcoming projects, the company is also preparing to implement the system from the planning and development stages.
Looking ahead, NeoWiz plans to enhance its capabilities by using AI to analyze user anomalies or difficulties encountered during gameplay, supporting game operators in their decision-making. If a user consistently struggles at a specific point, the system will help identify the cause of the hurdle and determine if difficulty adjustments are necessary.
Kim stated, “For instance, if a specific user continuously loses resources in games like Matgo or poker, we can analyze the reasons and reflect that in compensation methods or update directions, ultimately helping to reduce user drop-off. The goal of utilizing data is to create better games and ensure users have a more enjoyable experience.”
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.