Helping consumers make intelligent fashion choices
Employees can access and share data for internal decision-making
Yearly increase in data volume supported
“Databricks Data Intelligence Platform has been pivotal in our transformation journey. Now, our whole organization, from data experts to nontechnical users, can effortlessly utilize data in all aspects of our business to redefine the online shopping experience.”
Retailers today are increasingly adopting analytics and AI to take advantage of explosive growth in business data to increase customer loyalty and conversion. MUSINSA, an online fashion platform in Korea with 13 million subscribers, recently underwent rapid growth in various aspects of its business, including new subscribers, sales, and the number of brands on its site. As a result, the business faced increased data and complexity, prompting the implementation of the Databricks Data Intelligence Platform to address diverse data needs arising from this increase. With consumer and sales data at its fingertips, it can make smarter decisions, introduce recommendation services, and streamline operations to maintain its position as the No. 1 fashion store in Korea.
High growth requires enhancement of existing data platform
MUSINSA offers products from more than 7,000 street brands, global luxury brands, and designers to millions of consumers across Korea. The company also produces MUSINSA Magazine, which delivers content on the latest fashion trends in Korea and abroad. MUSINSA is increasing its competitiveness by launching its own homegrown labels, leveraging data and its industry expertise.
MUSINSA’s strategic vision is to offer personalized recommendations for customers, exposure to a wider group of store brands, and enhanced operational efficiency through data, analytics and AI. To achieve this, they sought to build an accessible data environment that facilitates data-driven decision-making.
However, before creating a new data environment, the company needed to find a better way to manage the recent rapid growth in data. MUSINSA has grown to include 7,000 partner brands, over 13 million members, 4 million monthly unique users, and an annual transaction volume of KRW 2.3 trillion in 2021. As a result of this growth, over the past 4 years, MUSINSA experienced a 40%–60% annual increase in the amount of data it was capturing on its platform and through its data users and diversified services. This includes both structured data — such as sales figures, subscriber numbers, product details, user reviews and behavioral logs — as well as unstructured data such as product and review images. The company currently processes close to 5 billion data points monthly, encompassing both transaction data and user behavioral logs.
“As the business grew, we wanted to integrate distributed services into a single data platform and configure different environments based on our employees’ need for data across different personas,” said Hwansung Yu, Team Lead of Data Platform at MUSINSA.
Managing access privileges by employees with various roles and jobs using Unity Catalog
To provide nearly 1,500 employees with a more accessible data environment catering to their varying levels of data usage and work requirements, MUSINSA implemented Databricks Data Intelligence Platform as the centralized platform to store, understand and analyze all data types. MUSINSA selected the solution because it integrates with the company’s existing Amazon Web Services (AWS) cloud environment, offers support for open source technology, and provides centralized data governance capabilities.
MUSINSA needed to establish data access and usage guidelines for a diverse range of data users, including general users, data analysts and data scientists, to ensure an accessible and versatile data platform. With Unity Catalog, they can take advantage of a unified governance model, while also implementing tighter access control. “We considered different options that could support an environment where not only data experts, but also users with no data skills, can easily access and utilize data. We were able to solve our challenges with Unity Catalog,” said Yu.
With Unity Catalog, MUSINSA can easily manage user permissions and seamlessly visualize and monitor data flow through data lineage for all SQL and Python workloads.
Accelerating the development of an accessible data platform
In just 3 months, the MUSINSA data platform team integrated all data operating on the AWS cloud into the Databricks Data Intelligence Platform. The data migration was completed smoothly because of Databricks’ seamless and flexible integration with AWS. Additionally, because Databricks Data Intelligence Platform supports open standards and technologies, most of the existing code that processed data based on Apache Spark™ could be reused, delivering a necessary level of extensibility within the engineering and data teams.
By using Databricks Data Intelligence Platform, MUSINSA provides the opportunity for all employees to access, share and prepare data for analysis. The MUSINSA data platform team also implemented strong governance via Unity Catalog, which allows granular access controls down to the data column level. “The key outcomes of our data platform are accessibility, governance and cloud integration,” said Yu. Leveraging its enhanced data capabilities, MUSINSA can make improved business decisions, facilitating seamless data management and the development of AI/ML models. Additionally, with faster access to data, the company’s employees are more productive.
Moving forward, MUSINSA intends to expand its use of Databricks Data Intelligence Platform, employing MLflow for ML lifecycle management and integrating existing AI/ML models into Databricks. Additionally, the company aims to develop an advanced data environment, empowering all users to make data-driven decisions to continually improve their customers’ online shopping experiences.
