Fashion retailer moves data out of the ‘lake’ and into the cloud
Harnessing customer-specific data is paramount for RueLaLa to engage shoppers and foster long-term customer relationships.
However, the road to uncovering the right information has not always been a smooth one. The membership-based e-retailer’s enterprise data warehouse stored all of its structured information, including how many shoppers visited its e-commerce site, where they were going and what they ordered — the foundation needed for marketers to “understand the sales funnel, according to Erick Roesch, RueLaLa’s director of business intelligence and data warehousing BI/DW, told Chain Store Age.
As the retailer connected with shoppers through more touchpoints, it collecting a new data stream — unstructured data. These details are related to raw, unorganized data that is often created through social media, and email and clickstream activity filtering through RueLaLa’s website and mobile app. Besides enters the company at very high volumes and velocities, this unstructured data began straining its enterprise data warehouse, making it “difficult to get insight into the upstream stages of the sales funnel,” he said.
To harness this swelling information, RueLaLa created data lakes, or repositories designed to hold vast amount of raw data. Besides creating a siloed “islands” of information that were disconnected from its enterprise data, these disparate repositories made it impossible to leverage information needed to create personalized email campaigns.
“Each interaction, or even answering one single question, became a complex engineering effort,” Roesch said.
The struggle pushed the retailer to create a next-generation database platform to house all of its data. The ideal solution was a cloud-based platform that could hold information from databases and data lakes, and scale as the velocity of data increased.
“A cloud-based platform would enable us to build out data storage without a large capital investment,” Roesch said. “This also would give us direct access to analytics unique to our needs.”
After choosing a platform from Snowflake, RueLaLa spent three months curating data from the data warehouse and data lakes in its new centralized repository. By the third quarter of 2016, “we had a complete view of our all data, from emails and click streams to traditional information,” Roesch added.
With data combined in one location, buying and merchandising teams can easily analyze previously viewed and purchased product — a move that improves merchandise buying decisions, as well as drives more accurately curated information online. And with insight into viewed merchandise, brand affinities, and past purchases, marketers have details needed to personalize emails.
“Unlike mass emails, our communications feature between five and eight products tailored to appeal to each specific member — a move that drives higher open and click-through rates,” Roesch explained.
The data also helps RueLaLa understand negative experiences.
“By applying analytics to understand churn rates, we can create new incentives to engage them, improve the shopping experience and close the sale,” Roesch said.
No comments found