Customer Case Study: Rue Gilt Groupe – Databricks

Rue Gilt Groupe

Customer Case Study

Rue Gilt Groupe

Rue Gilt Groupe strives to be the most engaging off-price, online style destination, connecting world-class brands with the next-generation shopper. Rue Gilt Groupe uses Databricks on AWS to power their recommendation engine that runs on their website. They also use Databricks and the embedded Tensorflow framework to do image comparisons to drive visual recommendations.

Vertical Use Case

Recommendation Engine

Technical Use Case

• Ingest and ETL
• Machine Learning
• Deep Learning

The Challenges

  • In order to give each of their 19M shoppers a unique shopping experience, they must ingest and analyze over 400GB of data per day.
  • Their data pipeline consists of 20+ different queries to transform the data for the data science team which was a highly computationally expensive process.
  • Scale was also a challenge as they lacked a distributed system to meet the performance goals they required.

The Solution

Databricks provides Rue Gilt Groupe with a fully managed, distributed platform that allows them to ETL their data much faster and leverage machine learning to power their recommendation engine.

  • Estimated query speedups of 60%.
  • Parallel model building and prediction resulted in a 25x time savings over competing platforms.
  • All of these improvements and benefits have made a tangible impact to their business, contributing to a 10x increase in purchase engagement, repeat visitation rates of over 60%, and incremental revenue gains of over 5% in a beta release program.

Databricks has enabled our data science team to work more collaboratively, empowering them to easily iterate and modify the models in a distributed fashion much faster.

Benjamin Wilson, Lead Data Scientist at Rue Gilt Groupe