TNL Mediagene Builds a Data-Driven Content Platform with MongoDB Atlas

TNL Mediagene is a technology and digital media company providing AI-driven advertising, marketing technology, content commerce, and data analytics solutions, and operating multi-language digital media brands across Asia, with operations in Taiwan and Japan. Its portfolio features leading Chinese, Japanese, and English-language media brands, including The News Lens, Cool3C, Roomie, and Business Insider Taiwan. Widely recognized for its diverse content and in-depth reporting, the group is redefining media and news services in the digital age.

TNL Mediagene stated that the group has consistently invested in digital capabilities to power data analytics, membership management, and content integration. The company’s goal is to redefine content value through technology—turning data from a by-product of news reporting into an engine for business growth.

As the number of media brands operating under TNL Mediagene expanded, and content volume surged, the company’s IT infrastructure couldn’t support the scale and agility required. TNL Mediagene turned to MongoDB Atlas’s fully managed cloud database service to move away from this legacy architecture and build its new-generation data-driven content platform: Inkmagine.

Building on MongoDB, TNL Mediagene realized its vision of ‘develop once, deploy everywhere’, enabling content creators, editors, and advertising and business teams to collaborate within a single environment.

Slow search speed and maintenance burden drive digital transformation

In its early years, TNL Mediagene relied on a relational database for member login, article management, and advertising modules. As data volume surged, search performance began to slow down. To address this, the team in Taiwan deployed an Elasticsearch instance, which improved query speed but created a heavy, costly maintenance burden.

In addition, each time a new media brand was added to TNL Mediagene’s portfolio, the team had to redeploy an entire backend system, making data integration difficult. Different media brands require highly varied data formats—for example, tech media need a “product model” field, while lifestyle outlets require “location” or “author bio” fields. TNL Mediagene’s relational database made it difficult and time-consuming to adjust schema structures, which also led to downtime.

Furthermore, the company’s expansion into Japan and Hong Kong showed the limitations of a single-region database architecture. Overseas visitors had to connect to servers in Taiwan, resulting in high latency. Traffic spikes driven by breaking news or major events could increase load over tenfold, requiring manual scaling to avoid downtime.

Addressing these challenges by building a more flexible data-driven content platform was critical to the group’s transformation.

MongoDB Atlas powers TNL Mediagene’s data-driven platform

TNL Mediagene evaluated database solutions based on four critical needs: faster search performance, reduced data complexity, support for the company’s expansion into multiple markets, and management of Site Reliability Engineering (SRE) services.

MongoDB Atlas met all these requirements with its on-demand scalability, integrated data analytics, search, and multi-cloud deployment capabilities. MongoDB also solved the main  challenges that came with using a relational database, such as performance limitations, low resiliency, high maintenance, and an inability to support global deployments seamlessly.

TNL Mediagene implemented MongoDB Atlas Search with built-in full-text search to replace its self-deployed Elasticsearch instance. This significantly improved the search speed for articles and media assets while eliminating operational burden. The team also brought in MongoDB Atlas Vector Search, which powers AI-generated image feature vectors for semantic similarity search, even when images do not have captions. This significantly improved the efficiency and accuracy of media asset searches.

For data structure management, MongoDB Atlas’s flexible document model enabled the Inkmagine platform to store diverse media formats and various article types, without schema overhauls, effectively reducing downtime risk.

Furthermore, as MongoDB Atlas supports major cloud platforms, including AWS and Google Cloud, TNL Mediagene could finally build an architecture that fits its multi-cloud and multi-region needs.  With data nodes distributed across Taiwan, Japan, and Singapore, global access latency was reduced to under 100 milliseconds. When breaking news drives sudden traffic surges, auto-scaling now automatically adjusts compute resources, setting maximum and minimum tiers to help control costs and maintain the platform’s stability.

Search performance increases 90%, operations simplified

Since launching Inkmagine, TNL Mediagene has achieved significant improvements in search performance and system operations. Article search speed has increased by 90%, while media asset search is now 50% faster. The platform’s simplified architecture has eliminated the need for additional databases, enabling the development team to maintain a single backend system for all brands.

This consolidation has drastically reduced the time required to launch new brands, shortening the process from several weeks to just a few days. With all data centralized in MongoDB Atlas, the analytics team can now conduct multidimensional analysis directly through Business Intelligence tools, enabling initiatives such as content recommendation and segmented member marketing. 

Looking ahead, TNL Mediagene plans to leverage machine learning and AI models to expand personalized content recommendations and advertising, further enhancing the platform’s commercial value.