Saks achieves data efficiency and enables AI with Fivetran

“Fivetran allowed us to reconstruct our entire data ecosystem in an unbelievably short time and enhance our technology infrastructure.”
- Mike Hite, CTO at Saks

Saks is an iconic brand in luxury retail. To meet consumer’s changing expectations, Saks has evolved into a digital-first company delivering highly personalized and elevated experiences to customers throughout their entire shopping journey.

Leading the digital transformation effort is Mike Hite, Chief Technology Officer at Saks. When he joined in 2021, the company was just beginning its transformation. Hite’s experience at high-growth companies like WeWork and Airbnb helped him see an opportunity for much greater speed and agility at Saks. At the time, Saks was burdened by cumbersome, slow ETL data pipelines that required constant attention and delayed data-driven decisions.

Saks needed a modern data stack to fuel innovation and agility. Hite brought in Fivetran, Snowflake and dbt, transforming Saks’ ability to perform and deliver on time to value.

Replacing legacy tech with the modern data stack in just 6 months

In 2021, Saks was established as a standalone e-commerce company with an exclusive partnership with the Saks Fifth Avenue stores and underwent a transformation to become a digital luxury pure-play. At the time, its data infrastructure wasn’t set up to most efficiently generate the results the company needed. They had dozens of custom ETL pipelines with no monitoring and faced major bottlenecks. 

Technical staff managed a complex enterprise service bus and a large infrastructure footprint. It took weeks — and sometimes months — to integrate new data sources. At this pace, Saks wasn’t set up to innovate as a luxury e-commerce platform and would be unable to quickly adapt to trends or deliver the online shopping experiences that customers expect in the fast-moving luxury e-commerce industry. Given its use of customer data, Saks maintains a customer data governance framework that helps the company ensure it is responsible with customer data and uses it to better serve customers. 

Hite joined Saks with a clear vision: to transform Saks into an agile, data-driven retailer backed by industry-leading technology. His first move was to bring in Fivetran to replace the existing ETL pipelines. On a single 40-minute call, he aligned his entire team by setting up a pipeline in Fivetran — a task they would have spent weeks completing. Hite laughs, “That one call was the end of the conversation. The direct time-to-value input is what landed the case for Fivetran.”

Time to value is one of Hite’s key metrics for success. With Fivetran, his team was able to onboard 35 data sources in 6 months, which Hite calls “a feat that would have been unimaginable with our old architecture." Saks can now onboard new data sources within hours, pulling in data from dozens of databases near real-time every 5 minutes. 

Data is ingested with Fivetran, stored in its raw form in Snowflake and transformed using dbt. This new architecture is changing the way Saks does business, and it’s having a positive impact  on the customer experience.

“This is the beauty of the data ecosystem we’ve built with Fivetran as a core piece: We can think about data in a fundamentally different way, less consumed with the cost of getting it in and instead focusing on the value it can bring to our customers and brand partners.”
- Mike Hite, CTO at Saks

[CTA_MODULE]

Beating high expectations with high tech

Saks’ focus on building high-touch customer relationships is a key component of its success. Whether before or after a customer’s purchase, Saks is purposefully establishing and fostering meaningful relationships. It’s why customers have come to expect elevated shopping experiences and services, including throughout their post-purchase experience. 

When customers contact Saks, customer service agents receive data-driven recommendations from large language models (LLMs) to enhance every caller’s experience. Saks uses several tools for real-time natural language processing (NLP) and understanding call sentiment. The investment is paying off, notes Hite, “We tweaked our model to where we could provide a summary and actionable insights into the customer experience for agents who have high touch with our customers.”

In a time when most retailers use data to simply fuel analytics needs, Saks is deploying data products into segmented Snowflake data marts. Each mart offers specific types of data to serve data engineers, data scientists, BI analysts, vendors and brand partners. Hite reports the marts are “extremely successful” and help Saks to deliver valuable insights to our vendors and brand partners in new ways.

Data is powering both internal and third-party apps, providing consistent and fresh insights across Saks and its vendors and brand partners. With the modern data stack, Saks can focus on building core KPIs across the business, like ensuring packages arrive on time. 

“Our ecosystem of Fivetran, Snowflake and dbt allows us to do in 2 weeks what could take 6 months to do. We can experiment, try things and scale up quickly to take advantage of opportunities as they come.”
- Mike Hite, CTO at Saks

Saving significant costs per year staying lean with Fivetran and the modern data stack

The modern data stack brings tremendous value in staffing, productivity and overhead costs. For Hite, it’s simply another expression of time to value and business impact. He has a handful of data engineers handling all ingestion across the entire enterprise. “The previous structure that supported Saks required 4-5X the current team. Now, fewer data engineers are much more efficient because of Fivetran.”

This is a major shift from how the Saks team looked years ago. Their custom ETL jobs required many more engineers to manage, including the infrastructure, enterprise service bus, data transformations and data movement. Custom pipelines often came with technical debt and were built by engineers with specialized knowledge. 

Under Hite’s leadership, his team has transitioned out of manual, slow work and into high-value activities like in-depth analysis and adopting new technologies like AI, ML and GenAI. His approach has enabled his team to provide exceptional experiences supported by a modern data ecosystem. “The beauty of Fivetran is that it solves a very complex problem very simply for us: ingesting lots of different data. It’s one of the fundamental pieces of our AI strategy and allows us to bring in new novel data sets and determine whether they’ll be useful for us.”

Hite continues, “We love Fivetran and can’t rave about it enough.”

[CTA_MODULE]