The Major Domestic Merchandise Retailer operates a nationwide chain of retail stores. The Company, through its retail stores, sells a wide assortment of merchandise principally including domestic merchandise and home furnishings, as well as food, giftware, health and beauty care items, and infant and toddler merchandise.
The Major Domestic Merchandise Retailer kicked off two big-data initiatives. The first business case was to bring in clickstream Omniture data from Adobe into the data warehouse so that the customer analytics team could understand online behavior tied to their website. In the existing state, the data was sent on request by Adobe which led to long delays in data acquisition. Secondly, the warehouse team required a platform to house low latency and seldom queried data that was clogging up their existing Netezza and Teradata systems.
iOLAP architected a solution to build out a Hortonworks Hadoop cluster to serve as a data lake and ingestion platform for incoming enterprise data. Hive was utilized to store daily clickstream data which was then parsed and sent to Netezza. Using Netezza UDFs, the clickstream data was further parsed into data that could be tied to customer activity around college searches, product and event lists, and visitor keywords. Further, Hadoop’s HDFS was leveraged as an archival location for the data that was no longer required to exist on Teradata and Netezza.
Leveraging Hadoop, Major Domestic Merchandise Retailer enabled the ability to analyze clickstream data as well as laid the groundwork for a cheap storage alternative in the form of a data lake.