INTRODUCTION
Based in Hong Kong, HK01 is a premier digital-native media platform that delivers dynamic, real-time content to Chinese-speaking audiences worldwide, and personalized services through various digital channels. Consistently ranked as a top ten local internet company, HK01 is one of the largest Chinese-language media platforms globally.
Since its launch in 2016, HK01 has gained strong traction and influence in Hong Kong. In 2022 alone, the platform accumulated close to 10 billion page views, and recorded over 1.4 billion page views and more than 17 million active users in a single month. In less than a decade, HK01 has reached far beyond Hong Kong, connecting with Chinese-speaking audiences worldwide—including in Taiwan, the United States, Canada, and the United Kingdom.
Initially built on a SQL database, HK01 quickly encountered challenges and limitations which led the company to transition to a modern document database architecture. Migrating to MongoDB Atlas empowered HK01 to build an architecture that was much more flexible, and able to scale.
MongoDB has been instrumental in enabling HK01’s tech team to innovate and rapidly deliver dynamic digital capabilities to a global market. It has also set the foundation for the next stage of the company’s growth, which include deepening HK01’s search capabilities with products like MongoDB Atlas Search, as well as geographic expansion to the Greater Bay Area in Mainland China.
THE CHALLENGE
Scaling a growing digital media platform with rich functionalities
HK01’s digital platform was originally built as a monolithic web application using PHP and MySQL.
“As we expanded into a broader internet lifestyle service platform, we recognized the need to modernize our technology foundation to support scalability, agility, and rapid feature development,” said Zhaozhou Li, Senior Product Director at HK01.
HK01 was looking for a modern database platform that could solve a wide range of challenges:
- Scalability: The company’s rapid increase in content and user interactions highlighted the limitations of HK01’s existing technology stack, which was initially built with PHP and MySQL. The monolithic architecture could not efficiently handle the company’s massive scale—HK01’s high-traffic platform needs to serve over 10 billion page views annually—frequently leading to bottlenecks in performance.
- Evolving content and multi-model data needs: Transitioning into a comprehensive internet lifestyle platform required a dynamic and adaptable data infrastructure. Furthermore, with HK01 regularly launching interactive content formats such as interactive FAQs, interactive maps and interactive tables, it needed a more flexible database model. The team needed to be able to add or modify content structures—like new metadata fields or engagement metrics—without downtime or heavy schema migration. It also needed a database that could support diverse and rapidly changing data requirements without extensive refactoring.
- Operational complexity: With a primary focus on delivering high-quality user experiences, managing database operations was proving to be a significant drain on engineering resources. The lack of dedicated in-house database administrator expertise made maintaining high availability and optimal performance challenging without impacting development cycles. HK01’s team spent significant time scaling related incidents and tuning the database to fit the traffic needed.
- Real-time user engagement: Delivering personalized, real-time content experiences demanded an agile technology environment that could support fast-paced innovation and deployment. This was difficult to achieve with HK01’s rigid, legacy relational system.
Compliance and latency management: Compliance with local data sovereignty laws and low-latency service delivery across regions was a critical requirement for HK01 to operate, as well as to expand into the Greater Bay Areas in Mainland China. This puts extra demand on the data platform.
“We needed a database solution that could evolve with our product,” said Dick Tang, Director of Development Operations at HK01. “The traditional relational database model was too rigid for our rapidly changing requirements. On the other hand, MongoDB’s flexible document model allowed us to iterate quickly, adapt our data structures on the fly, and launch new features without costly schema migrations.”
THE SOLUTION
The power of a fully managed platform, with native search functionalities
Shortly after launching, HK01 realized the limitations of the SQL model. Consequently, HK01 embarked on a journey to re-architect its platform on MongoDB Atlas.
Rather than attempt a single ‘lift and shift’ large-scale migration of legacy data, HK01 adopted an incremental approach. New services and features were designed from the ground up using MongoDB Atlas. Concurrently, legacy components continued to operate alongside them, and over time HK01 gradually deployed the new microservices. Each microservice was isolated and independently scalable, which allowed HK01 to shift specific workloads and data models to MongoDB Atlas in a controlled and manageable way.
“This phased rollout enabled us to minimize risk and avoid major disruptions,” said Li. “It also allowed our engineering teams to learn and adapt MongoDB best practices as they progressed, making the transition smoother and more sustainable.”
“We chose MongoDB Atlas because its document-based model aligned perfectly with our need for flexibility, speed, and scalability—especially as a digital-first company operating in a fast-changing media and lifestyle space,” said Tang.
The company is also currently migrating to MongoDB Atlas Search to employ a fully integrated, scalable search solution. Search plays a critical role in HK01’s platform; the editorial team needs the ability to quickly retrieve related content for content creation, curation, and real-time news coverage. Concurrently, the user-facing side of the search platform is a key entry point for content discovery.
HK01 tried several solutions and products to help enable this, but none were able to provide the performance, search accuracy and system simplicity needed.
“One of the key benefits of MongoDB Atlas Search is that it eliminates the need to manage a separate search infrastructure, allowing us to reduce operational complexity and focus on improving the search experience itself,” said Tang.
OUTCOME
Exceptional reliability, built for innovation and expansion
“One of the most significant benefits of MongoDB Atlas is its exceptional reliability,” said Tang. “Since adopting Atlas, we’ve experienced zero downtime related to MongoDB cluster failures, which has been critical in maintaining the trust and confidence of our users.”
HK01’s services need to remain stable and accessible around the clock, even during traffic spikes or system upgrades. MongoDB Atlas’s built-in high availability, automated failover, and resilient infrastructure have been instrumental in solving this challenge.
And from an operational perspective, MongoDB Atlas’s architecture allows HK01 to iterate rapidly without being constrained by rigid database schemas. Using MongoDB Atlas saves HK01 team’s time when it comes to troubleshooting for scaling related incidents—developers can focus their time on coding, building features and delivering value to end users.
“We can evolve our data models incrementally, test new ideas quickly, and pivot when needed,” said Li. “MongoDB not only supports our product evolution, it also strengthens our engineering culture and our ability to build and retain a high-performing team.”
MongoDB Atlas Search offers a rich set of tunable full-text application search parameters. This gives HK01 fine-grained control over search relevance, indexing strategies, and query behavior. This flexibility is particularly important for HK01, which supports Chinese languages.
“Choosing the right tokenizer dramatically impacts search accuracy and user satisfaction in these languages,” said Tang. “With Atlas Search, we’re able to experiment with different analyzers and tokenization strategies to deliver more relevant and intuitive results to both editors and readers.”
Thanks to MongoDB’s flexibility, HK01 is confident that it can keep growing, particularly as it expands into Mainland China, where it will have to comply with strict local data sovereignty laws.
.svg)