Product descriptions:
Cafe24 is the number one online shopping mall platform in Korea, supporting more than 2 million brands as they build, operate and scale their businesses globally. At the scale of Cafe24, data sits at the center of commerce innovation. Trillions of won in transactions and hundreds of gigabytes of data generated daily create both opportunity and complexity for AI-driven growth. Before Databricks, Cafe24 struggled with fragmented systems, inconsistent KPIs and slow, manual data processes that limited automation and made it difficult to apply AI at scale. After unifying data on Databricks, Cafe24 connected analytics directly to execution, standardizing insights, enabling autonomous workflows and turning operations into a continuous learning loop. What is now possible is an entirely new operating model where data-driven actions are automated and insights are accessible across teams in real time. Brand growth can be repeatedly reproduced, unlocking faster experimentation, global expansion and measurable gains in customer acquisition and order value.
Fragmented legacy data environment limits brand growth
Today, barriers to entry in the e-commerce market have lowered, but achieving sustainable growth has become more difficult. Differentiation through price competition alone has clear limits. As businesses are expected to operate across multiple channels, expand globally and continuously elevate customer experience, brand competitiveness is ultimately determined by how quickly an organization can learn and execute using a data-driven operating system.
As businesses are expected to operate across multiple channels, expand globally and continuously elevate customer experience, brand competitiveness is ultimately determined by how quickly an organization can learn and execute through a data-driven operating system.
To help brands compete, Cafe24 has evolved beyond a simple online store-building tool into a commerce platform that supports brand growth end-to-end. From storefront creation to marketing, global expansion and sales, Cafe24 covers the full e-commerce lifecycle. Backed by an ecosystem of approximately 6.2 million members and KRW 11 trillion in annual transaction volume, Cafe24 has the scale to truly move the market. The company launched Korea’s first D2C model, connecting global social media and online stores.
At the core of Cafe24’s support for brand growth is Cafe24 PRO. Cafe24 PRO is a professional operations service that supports the full scope of online business, from store setup to product and inventory management, order and payment processing, marketing and global expansion.
However, as services expanded rapidly, data became distributed across multiple systems. As customer behavior, operational actions and marketing performance data continued to grow, the legacy data environment came under increasing pressure. With more than 300GB of log data generated daily and over 700 tables to manage, processing and operational efficiency became increasingly constrained.
KPI definitions and analysis methods differed by organization and domain, making it difficult to analyze and leverage data consistently across the enterprise. Even for the same metric, differing aggregation rules led to inconsistent interpretations. In addition, critical data such as operational actions, outcome metrics and service logs was siloed, significantly increasing the time required to identify root causes and assess impact when issues occurred.
The legacy environment also had limitations in scalability and automation. Data ingestion, cleansing and validation required substantial time, extending decision-making lead time. And since stable training datasets and governance foundations were not sufficiently established, bottlenecks frequently emerged in the data preparation phase when applying AI and machine learning.
To address these structural challenges and improve the efficiency of data-driven execution, Cafe24 needed to unify data into a single system and connect analytics to operations in one continuous flow. “To realize the ‘repeatable operations that reproduce performance’ that Cafe24 PRO aims for, we needed to fundamentally reorganize the underlying data system that supports it,” explained Hyuncheol Jang, Senior AI Engineer. Cafe24 chose the Databricks Data Intelligence Platform as its enterprise-wide data system to implement standardization, governance and consistent operational automation at scale.
Supporting an integrated e-commerce data intelligence platform on Databricks
Cafe24 applied Databricks across Cafe24 PRO operations, transforming key business domains into a data-driven execution model. Based on customer purchase patterns and behavioral data, Cafe24 refined promotions by conversion stage such as sign-up, first purchase, repeat purchase and churn prevention. This approach enabled the company to design and optimize execution items including cart reminders, revisit triggers and benefit timing optimization.
Cafe24 also enhanced the purchasing experience by using data to improve product information and merchandising strategies. By making trusted data accessible across teams through the Databricks Platform, Cafe24 enabled faster experimentation and decision-making. These initiatives were continuously iterated through performance measurement and feedback loops, reinforcing a cycle of ongoing optimization.
In addition, Cafe24 increased search-driven traffic through SEO optimization. The organization also simplified complex channel operations into a single operating system and improved overall business efficiency by structuring global store launches and operations. These improvements were enabled by an integrated data foundation built on Databricks’ data lakehouse architecture.
Cafe24 consolidated large volumes of data, including transactions, operations, customer behavior and service logs, into a single lakehouse. A Medallion architecture was applied to manage data quality and reliability in a structured way. By leveraging multiple workspaces and clusters, Cafe24 scaled data environments by organization and workload. Centralized governance, metadata management and fine-grained access control were handled through Unity Catalog, allowing teams to scale without sacrificing control.
The Databricks Platform provided an enterprise-wide data environment that business users could access and analyze using natural language. This enabled non-technical teams to participate directly in data observation and analysis.
A defining feature of this new Databricks environment is its shift toward operations as a sustainable learning loop. Cafe24 standardized core KPIs, established hypotheses to explain metric movement and executed those hypotheses as verifiable operational actions. Outcomes are continuously observed and fed back into the system, reinforcing a cycle of ongoing learning.
This approach is powered by Agent Bricks, Databricks’ multi-agent AI framework. Using Agent Bricks, complex analytical tasks are decomposed and orchestrated across specialized AI agents that collaborate on reasoning, analysis and execution. This unifies data exploration, analytics and operational action into a single end-to-end workflow.
Cafe24 successfully built an operating environment where data drives execution and decision-making through Databricks. Operational teams can observe results immediately through BI dashboards, analytics apps and natural-language interfaces, then quickly translate insights into action. According to Hyuncheol, “The platform we built on Databricks made it possible to run operations where performance can be repeatedly reproduced.”
Providing better brand insight with efficiency gains from autonomous workflows
With the Databricks Data Intelligence Platform being the system that underpins Cafe24 PRO, there has been a significant improvement in operational efficiency and business outcomes. Databricks replaced a labor-intensive model in which single-store operators manually performed up to 100 repetitive tasks per day. Now, Cafe24 PRO proactively executes these tasks, substantially reducing the operational burden.
Brands using Cafe24 PRO showed an average transaction value distribution about 40% higher than general shopping malls of similar scale. Over a six-month period, Cafe24 PRO recorded strong results: a 24% increase in total revenue, a 64% increase in average order value and a 36% increase in new customers.
In marketplace operations, manual work that previously took more than three hours per marketplace was reduced to under five minutes through a single login and click. For global stores, Cafe24 also demonstrated rapid expansion outcomes including launching a global mall in just 20 minutes and achieving first overseas sales within 10 days. This significantly shortened the time to market for global expansion.
Building on these outcomes, Cafe24 plans to continue evolving the PRO platform within the Databricks environment with a strong focus on advanced AIOps. The organization will enhance automatic model drift detection and retraining while strengthening data quality monitoring and self-healing mechanisms to improve operational stability. There are also plans to introduce adaptive training infrastructure and a streaming-based training data architecture to further elevate analytics and AI capabilities.
In addition, the company will expand autonomous execution and operational optimization using Databricks Serving Endpoints and Serverless Compute. These efforts will be supported by workflow-based GitOps along with stronger cost and security governance to enable sustainable data and AI operations. “The core of PRO is to standardize high-performing operational methods through data and rapidly apply them to each brand—reproducing brand growth by combining data with an execution system,” said Miji Jeong, Director and Head of Cafe24 PRO (FGS Team).
Through collaboration with Databricks, Cafe24 will continue advancing an operating model where data directly translates into execution. With this integrated data foundation in place, the organization will remain a leader in sustainable brand growth for an even wider range of brands in Korea.