Implementing Refined Marketing at Low Costs with Game Publishers
Step up the digitalization of your business with Alibaba Cloud 2020 Double 11 Big Sale! Get new user coupons and explore over 16 free trials, 30+ bestselling products, and 6+ solutions for all your needs!
By Lu Yang (Tanjian)
Editor: Fan Ao (Duanmu)
Industry Trend — Refined Marketing of the Game Industry
The COVID-19 pandemic in early 2020 led to a substantial increase in the number of users in the game industry. Servers of some products were unavailable because of too many gamers, and the stock prices of game companies in the secondary market began to soar.
The entire gaming industry was affected by various policy changes and suffered a severe setback two years ago, particularly in China. In 2018, the overall gaming and entertainment sector fell by 36.66% and struggled for most of the year. It wasn’t until December 2018 that China lifted the ban on approving new game titles, and the game industry began to gradually recover. Coupled with the recent developments in 5G, technical bottlenecks of cloud games and AR/VR are expected to be solved, and the future development of the game industry is looking brighter than ever.
The whole game industry has experienced an overall decline and then gradually recovered. Many game companies are facing problems, such as high promotion costs, difficulties in attracting users, and high traffic costs. Gamers’ high requirement for quality has led to growing competition among games and game companies. Therefore, game companies need to provide better games and do better marketing for better performance. With the continuous increase in traffic costs, the marketing of the game industry began to develop towards refined marketing.
Enterprise A is a unique and innovative game company that currently has more than 100 million mobile gamers worldwide. Its flagship mobile game became recognizable nationwide and eventually achieved great success worldwide. In recent years, Enterprise A has continued to develop innovative mobile games with interesting gameplay. Enterprise A has also committed to developing a game distribution business and bringing more playable games to gamers worldwide. Moreover, Enterprise A plans to provide high-quality platforms for major game developers to develop and release games worldwide. Enterprise A is currently the #1 mobile game publisher in China.
The Rapid Growth of Data Volumes and Demands of Analysis
Competition in the gaming industry is fierce. The rapid growth of data volumes and demands of analysis have brought challenges to the scalability, usability, and real-time capability for self-built systems. The game operating platform is required to achieve refined operations with real-time feedback and predictions. To implement real-time analysis, the advertising big data analysis department of Enterprise A took Hadoop as the core of its ecosystem and built its own big data system. However, with the rapid growth of data volumes and the increased business requirements for data analysis, problems in the following aspects of this big data analysis system were gradually exposed.
- Scalability: The growth curve of data volumes is high, making it difficult to scale up the Internet Data Center (IDC).
- Usability: The cost of self-maintained systems with Hadoop, Hive, and Presto combination is too high in learning and maintaining.
- Real-Time: Due to the increasing real-time requirements of busineses, Presto, as a real-time computing engine for direct queries, fails to meet expectations. The real-time performance of data analysis is also not good enough. It is still insufficient in real-time performance, even if the direct queries are output from high-performance databases after precomputing.
- Price-Performance Ratio: To ensure performance and stability, the cost of self-built clusters will greatly increase as IDCs scale up.
Low-Cost Solutions to Address Challenges of Growing Business Volumes
To address preceding business challenges and pain points of the architecture, the big data team of Enterprise A began to investigate the architecture and product selection. Enterprise A tried several open-source and distributed analysis engines and big data products successively. However, these engines and products all failed to meet the requirements of real-time and relational queries. Ultimately, the big data team decided to use Alibaba Cloud’s AnalyticDB for MySQL for Proof of Concept (POC).
AnalyticDB is an exclusive product that provides a real-time and online analytical processing (OLAP) service for massive data at high concurrency. It can analyze hundreds of billions of data records from multiple dimensions within milliseconds and provide data-driven insights into the business. In addition, AnalyticDB has many good features, such as speed, high flexibility, usability, scalability, and concurrency. After experiencing these features, the big data team of Enterprise A and Alibaba Cloud’s ApsaraDB Team worked together to build a next-generation platform for real-time operation and analysis for game advertisements based on AnalyticDB.
In this solution, attributed data is stored into storage-intensive AnalyticDB after logstash, and then stored into computing-intensive AnalyticDB after pre-processing for front analysis. This solution can replace the original system with Hadoop, Hive, and Presto. The new platform based on AnalyticDB has many advantages:
- Fast Querying Speed: The querying speed is more than 10 times faster than online transaction processing (OLTP), and several times faster than Presto. QPS of the new platform can be changed from the hundreds level to the tens of thousands level.
- Elastic Scaling: Nodes and configurations can be flexibly upgraded anytime as data volume increases.
- Easy-to-Use: Almost no change cost is incurred during migration from Presto, and most MySQL migration statements are compatible.
- Massive Scale: Nodes can be scaled up to thousands dynamically and linearly, supporting the storage and analysis of massive data.
Performance Improvements and Savings for Customers
By combining PolarDB with the high-performance AnalyticDB with large storage space, Enterprise A has created a next-generation platform for real-time data operation and analysis in the game buying market. Efficient game data operations have been achieved through cloud-native data processing and the analysis loop.
Users can mine the value of data by improving real-time data analysis. This can also help users develop their business. After building the next-generation platform, the analysis performance has improved five to ten times over. The platform greatly improves the business experience and promotes the launching efficiency transformation of the buying market.
The log tables of gamers’ behavior are growing, with over 100 million data records increasing daily. Instances of storage-intensive AnalyticDB are used for storage and analysis, which effectively reduces the overall use-cost by up to 75%.
With five to ten times better performance, 75% cost savings, and a great price-performance ratio, the real-time data operation of next-generation games will step into a new era.
