Connecting people with brands that bring them joy
Myntra uses Databricks to personalize every stage of the shopping journey
Reduction in overall data infrastructure costs
Faster real-time pipelines

Myntra, part of the Flipkart Group, is one of India’s leading fashion e-commerce platforms for fashion, beauty and lifestyle brands, offering over 3.5 million styles from 9,500+ brands. Striving to deliver a holistic shopping journey that engages customers at every step, the company recognized an opportunity to enhance their traditional data lakehouse to better support their AI ambitions.
Myntra revamped their architecture using the Databricks Data Intelligence Platform. This transformation has empowered them to centralize data, analytics and AI while enhancing scalability, efficiency and cost-effectiveness. Now, they can build real-time data pipelines that feed machine learning models that deliver a superior pre- and post-shopping experience.
The need for the legacy data architecture to evolve to keep up with business growth
As a key player in the fashion ecosystem, Myntra integrates technology across their operations, from design to delivery. However, their previous architecture couldn’t support new use cases at scale. According to Raghu Ram Vengalil, Myntra’s Director of Engineering, the company’s data warehouse experienced significant performance issues. “We were dealing with asynchronous file locking, which caused frequent failures and required multiple retries in our Spark jobs,” Ram explained, highlighting the barriers to insights in their old system.
Myntra’s data platform processes tens of petabytes of data, including billions of clickstream events daily and several hundred thousands of user queries. The complexity of their legacy architecture led to duplicate sources of truth, high maintenance costs and an inability to handle growing data volumes. Despite being designed to support scalability and seamless access across different teams, outdated infrastructure and management issues hampered Myntra’s data platform. These challenges, coupled with Myntra’s need for real-time data access and insights, drove the shift to a new platform.
To integrate technology, innovation and fashion, Myntra needed a unified data intelligence platform. Ram stated, “Our tech ecosystem relies on a strong data and ML platform to ensure timely delivery of quality data to our business analysts and seamless integration with engineering systems from storefront to supply chain.” Seeking improved scalability, faster time to market, lower total cost of ownership (TCO) and enhanced collaboration, Myntra was on the search for a modern data platform that could meet their evolving requirements.
Unifying and simplifying real-time data processing with Databricks
Myntra chose to transform their data architecture with the Databricks Data Intelligence Platform due to its unified lakehouse architecture and advanced capabilities. With Databricks, Myntra achieved seamless integration of batch and streaming data processing. Implementing Delta Lake, Myntra adopted a medallion architecture comprising Bronze, Silver and Gold layers, systematically refining raw data for better management and analytics.
Delta Lake centralizes semantics between batch and streaming operations, simplifying Myntra’s data architecture and significantly reducing data ingestion time. “We switched to a more Kappa-based architecture with Delta Lake, unifying our processing across different compute engines,” Ram noted, emphasizing how Delta Lake addressed previous bottlenecks.
Delta Lake’s features, such as auto-compaction and compression, reduced storage costs, while its enhanced metadata and optimization capabilities improved query performance. This accelerated speed allows data scientists and analysts to gain timely insights without compromising data quality. Furthermore, Databricks’ support for open source tools enables Myntra to leverage the latest technologies while maintaining flexibility in data processing and analytics workflows.
Overall, the Databricks Platform has enhanced collaboration across data teams, improved real-time analytics and delivered enhanced data governance capabilities. This transformation has empowered Myntra to maintain their competitive edge with the insights necessary to fuel personalized shopping experiences online and within the Myntra app.
Advancing ML use cases for personalized customer service
The adoption of the Databricks Data Intelligence Platform significantly impacted Myntra’s operations, directly improving their ability to deliver a seamless customer experience. By consolidating data on Delta Lake and optimizing storage, Myntra reduced overall infrastructure costs by 35%, enhancing their ability to scale cost-effectively. Myntra’s real-time data pipelines have also seen a 25% performance improvement — translating into faster, more personalized customer recommendations, enhancing engagement and conversion rates.
By standardizing on the Delta Lake format, Myntra also improved data governance and quality, offering a unified view of their data for better decision-making. This streamlined approach facilitated improved month-over-month growth in new use cases, including ML deployments and enhanced feature engineering, driving continuous innovation.
The efficient processing of clickstream data and the ability to provide low-latency aggregates, such as real-time click-through rates and order metrics, enabled Myntra to refine user experiences for their 70 million monthly active users.
Ram concluded, “Ultimately, Databricks empowers Myntra to serve our customers better, optimize operations and maintain our leadership in India’s competitive fashion e-commerce market. We will continue to leverage and scale data and analytics on the platform, fulfilling new use cases regularly for consistent innovation.”