A data platform that’s the perfect fit
Uniting the magic of Coach, Kate Spade and Stuart Weitzman, Tapestry is a leading New York-based house of modern luxury accessories and lifestyle brands. In the retail fashion industry, the faster a brand can get the right products to market, the better it can serve its customers.
“Being customer-centric requires us to have a data-driven culture. We must ask ourselves how we can optimize the supply chain, product design, predictability and propensity of customers to buy our products. These are all data challenges.”
Fabio Luzzi
VP of Data Science and Engineering, Tapestry
Tapestry is brand-agnostic when it comes to managing its enterprise data. A central data engineering team safely and legally manages data for all brands and all regions, except where otherwise required by law. “This is by design, as we want to move together as a whole and still let Coach, Kate Spade and Stuart Weitzman be able to independently make decisions and serve their customers,” says Luzzi.
Yet the company’s legacy Hadoop-based data platform was hard to scale and consumed a lot of time and resources to maintain. Luzzi and his team began looking for a modern enterprise data platform that could match the speed and scale that they needed—and Snowflake was the right fit.
Story Highlights
Apollo, Tapestry’s customer analytics platform: Powered by Snowflake, Tapestry’s self-service customer analytics platform in Tableau delivers valuable insights like which products to strategically mark down during retail events.
Seamless data sharing with trusted partners: Current data sharing has enabled a deeper collaboration with partners and will unlock future supply chain optimization use cases.
Future Unistore adoption: Tapestry is looking to unify analytical and transactional data in Snowflake to provide any application with real-time customer transactional data.
Doubling data sources while reducing costs
To achieve its goal of customer centricity, Tapestry chose Snowflake as its modern enterprise data platform to support all of its brands. As Tapestry continues to exponentially generate more data, from clickstream data to web traffic, Snowflake’s separation of storage and elastic compute, Zero-copy Cloning and multi-cloud capabilities are key winning features.
“On a daily basis, Tapestry processes about 4 billion rows and runs over 100 major data processes,” says Muhammad Chaudhry, Head of Data Engineering at Tapestry. “We’ve roughly doubled our data sources, and yet we have achieved significant cost savings from our legacy solution, while maintaining high performance and minimal maintenance.”
2x
More data sources with robust cost savings
Snowflake was a strong fit to maximize existing skills at Tapestry. “We have many SQL users, which Snowflake excels at. And with Snowpark, we can use Python and other languages of choice. Snowflake expands our ability to build new applications that solve business problems at scale without significant investment of cost and time,” Chaudhry says.
A self-service analytics platform for rich promotion and customer insights
Since Snowflake eliminates data silos across Tapestry, the company has consolidated its technology stack. For example, instead of having Google BigQuery to process JSON files for website analytics, that data can now be easily connected to existing internal customer data and directly build data models in Snowflake. According to Lexie Ye, Head of Data Science Consulting at Tapestry, “By bringing on Tableau and other tools within the same Snowflake environment, we were able to build out tools like Apollo, our self-service customer analytics platform.”
“Snowflake saves us a lot of time, and we don’t need to stitch things together.”
Lexie Ye
Head of Data Science Consulting, Tapestry
With the Apollo platform, Coach identified top products suitable for strategic markdowns during a retail customer event and was able to identify how business increased due to exclusions on select products. Using these insights, the business can know when to push and pull back on promotions. Another team at Coach found they could uncover who was a new multi-channel customer and what drove them to make their first purchase in a retail channel.
Simplifying data pipelines with Snowflake data sharing
Tapestry uses Snowflake’s Secure Data Sharing to acquire new customers. “We are currently leveraging Snowflake’s native data-sharing capabilities with our partners,” Chaudhry says. “Sharing data through our traditional data pipeline from our legacy system took us around six to eight weeks and required at least 10 to 15 people to help orchestrate. With Snowflake, it took us less than half a day to set up and a few more days for validation.”
Chaudhry continues, “Managing infrastructure is not strategic to us. Building strategic data products for our brands that help them grow is what Snowflake enables us to focus on.”
“Managing infrastructure is not strategic to us. Building strategic data products for our brands that help them grow is what Snowflake enables us to focus on.”
Muhammad Chaudhry
Head of Data Engineering, Tapestry
Unlocking supply chain optimization through data science models
The supply chain analytics team previously had to manually pull data from many systems into Excel, a messy process that didn't ensure cleanliness or data governance. Now that all supply chain data stores have been migrated into Snowflake, business users will have Tableau dashboards that offer robust and flexible analytic capabilities they’ve never had before.
As Snowflake-powered data science models become more advanced, data sharing will play a larger role in unlocking upstream and downstream supply chain optimization. From bringing in data directly from the manufacturers to sharing data with major shipping partners, this will enable Tapestry to tackle complex business challenges, such as optimizing product development timelines.
Bringing OLAP and OLTP data together with Unistore
Now that Tapestry has established Snowflake as its foundational data platform, the company is looking to leverage advanced features, such as Snowpark and Unistore, to further customer-centric decision-making.
“We have been greatly anticipating Snowflake’s Unistore offering. Our goal is for Snowflake to be the single, standard enterprise data factory of Tapestry—not only using OLAP data, but also by bringing in OLTP data into the same database for analysis.”
Muhammad Chaudhry
Head of Data Engineering, Tapestry
According to Chaudhry, “Snowflake’s Unistore feature could unlock a future where we can provide any application with real-time customer transactional data. And with Snowpark, Python and other languages available, we can potentially develop applications directly on Snowflake through its native applications framework.” Tapestry is currently piloting Unistore in several small applications within its pre-production environment, such as customer data management and customer 360.
