Tapestry Gains 360-Degree View of Customers by Powering Data and Analytics on AWS

Tapestry is the parent company of three global luxury fashion brands: Coach, kate spade new york, and Stuart Weitzman. These brands operate a total of 1,500 stores globally. In addition to its brick-and-mortar stores, Tapestry offers online shopping; its digital sales account for a significant portion of its yearly revenue. To collect data across its operations, Tapestry built the Tapestry Data Exchange, which acts as the data backbone for all its brands, analyzing and storing data from nearly 100 of its systems. Data stored on the Tapestry Data Exchange includes customers’ shopping histories, preferences, and other customer insights. Tapestry takes customer data privacy seriously. Customer data it collects is managed and processed in accordance with regional, state, or local privacy regulations. Tapestry Data Exchange is also used to process General Data Protection Regulation and California Consumer Privacy Act requests it receives from its customers and help it to quickly respond to new privacy regulations globally and deliver a privacy-first-oriented data solution.

Prior to migrating to AWS, the company used on-premises infrastructure that consisted of hardware to power its Data Exchange, which created data silos across its operations. Because of the disparate nature of the company’s on-premises infrastructure, Tapestry had difficulties using all its data to its advantage. “Our databases and analytics environments were fragmented,” says Justin Bussen, vice president of data engineering and data solutions at Tapestry. “We had to push data between internal solutions, which took a lot of time, coordination, and overhead.”

In 2018, Tapestry initiated a company-wide digital transformation, starting with migrating its analytics to the cloud. After gaining experience using AWS, Tapestry decided to modernize its Data Exchange so that it could support rapid innovation and drive deeper insights about its customers. “Our company vision is to be more customer centric,” says Fabio Luzzi, vice president of data engineering and science at Tapestry. “We want to use our data more systematically to drive ML predictions and automation across the value chain.”

Tapestry’s Data Exchange consists of three pillars: a data lake, a massively parallel processing system, and an operational data store. The company worked to optimize the performance and availability of all three pillars by adopting a combination of cloud and fully managed AWS services, including AWS Lambda, which gives companies the ability to run code without thinking about servers or clusters. “We want to consume compute power,” says Muhammad Chaudhry, head of data engineering at Tapestry. “We rely on AWS to achieve high availability—that way, we can focus on developing products, which is strategic to our business growth.” Tapestry has also benefited from the scalability of fully managed solutions on AWS, effectively optimizing its compute costs.

To transfer data quickly and securely across its three environments, Tapestry uses Amazon API Gateway. Using this solution, the company delivers key insights about customers, such as their purchase histories, to its point-of-sale systems across 1,500 stores and 18 distribution centers globally. “The elasticity that we have achieved would not be possible without Amazon API Gateway,” says Chaudhry. “It horizontally scales to our operations.” On average, Tapestry fulfills half a million API calls daily, and around the holidays, the company has scaled to 1.5 million API calls a day, peaking at 10,000 calls per minute.

For storing its data, the company has built a data lake on AWS using Amazon Simple Storage Service (Amazon S3), an object storage service built to retrieve any amount of data from anywhere. This data lake houses both structured and unstructured data and is the basis for developing Tapestry’s ML models, which it applies to a wide variety of key uses. To build its ML models, Tapestry uses Amazon SageMaker, which gives companies the ability to build, train, and deploy ML models for virtually any use case with fully managed infrastructure, tools, and workflows. By working in a scalable framework, the company has deployed dozens of highly complex ML models to production, including an ML model that performs customer segmentation in near real time. Using these insights, Tapestry designs products that align with its customers’ tastes and predicts how much inventory it needs to send to its stores. “Our ML models help us optimize the supply chain,” says Josh Ainsley, senior director and head of data science at Tapestry. “We can work to make sure that we have the right product in the right place for the types of customers that shop there.”

By deepening its use of ML, Tapestry can also better understand customer behavior to drive sales and personalize the shopping experience to customer’s needs. For instance, the company built an ML model that identifies VIP customers who purchase an abundance of products across its brands. As a result, Tapestry’s sales representatives can recommend products that align with customers’ preferences and cultivate strong relationships with shoppers. Tapestry also uses its data to offer customized promotions to customers, such as birthday discounts. “We want to keep our customers happy and encourage them to return,” says Ainsley.

Having furthered its technical capabilities on AWS, Tapestry is working on creating ML models that measure the correlation between priority shipping times and repurchase rates. By creating these models, Tapestry hopes to boost its ecommerce sales. “Using AWS has definitely changed our perspective,” says Luzzi. “We are more data driven and customer centric than we used to be.”