A large online retailer selling branded apparel, footwear, and accessories to consumers.
Customer Segmentation & Recommendation Engine
- Delayed insight & complex risk driver identification process
- Expensive data engineering overhead
- High cost for risk misalignment
- Increased average order value & revenue
- Greater customer satisfaction & retention
- More optimized ad spending
A large online retailer sought to expand customer wallet share and increase revenue by segmenting customers and building a recommendation engine that leverages data across multiple company divisions and jurisdictions.
The retailer was challenged to access and model the necessary data. Datasets such as historical purchase transaction records and detailed product features, including brand, color, category, etc., were stored in separate systems and had heterogeneous schemas. They were also owned by different company divisions and contained sensitive customer information.
Traditionally, the retailer had to duplicate, move, align, and combine these disparate data sources before building analytics. Unfortunately, in this case, the two divisions would not share their datasets, rendering efforts to construct an AI product recommendation engine ineffective.
The Devron platform was deployed to the customer's environment within a few hours. Devron Satellites, proximate to data stored in Snowflake and to data within BigQuery, allowed the data to remain within the domains in which they reside. The respective data owners provisioned specific datasets and applied privacy controls where required.
Using our remote data science platform, Devron experts clustered the product and customer transaction datasets to group similar products based on various features and similar customers based on purchase and cart history. Then, they built a hybrid content-based and use-based recommendation engine model that considered individual purchase or cart history, clustered products, and purchases made by other customers with similar profiles.
By leaving the data in situ and not exposing the underlying personal information, the retailer overcame the data access limitations across divisions and built a model that incorporated more data for greater accuracy and better insights.
Devron enabled the retailer to offer tailored product suggestions, increasing average order value and revenue. As a result, they were able to deliver a better customer experience and optimize ad spend across digital channels.
Devron enables data science teams to realize enhanced outcomes, improved data science agility, and reduced overhead—all without moving or exposing data.
Increased Average Order Value & Revenue
Deploying the hybrid recommendation engine allowed the retailer to offer tailored product suggestions to their customers, increasing average order value and revenue.
Greater Customer Loyalty
By offering their customers individualized recommendations, the retailer was able to deliver a better experience to their customers, resulting in greater customer satisfaction and higher customer retention.
Optimization of Advertising Spend
Through the clustering of similar customers, the retailer was able to better segment their market to deliver more targeted ads—ultimately, better optimizing their ad spend across digital channels.