Business Objective
Our client is a leading US-based furniture store doing business via both online channels and physical outlets.
The client wanted us to build a direct in-house email campaign management system (fully automated, flexible, and extensible) to identify target customers for each campaign and customize the offers/mailers that are sent out.
Challenges
- Data had to be extracted from traditional mainframe systems
- Dynamic rules engine framework had to be built to enable the business to add custom filters at any point
- Historical data of executed campaigns were not readily available
Solution MethodologyÂ
- Leveraged data ingestion pipeline to ingest customer and campaign data (including third party data on customer credit scores) and developed a data ingestion framework in Spark to parallelly extract multiple data from the mainframe systems
- Identified customers for address standardization based on customer profiling
- Stored all the data in the Hadoop data lake (generated by filtering/suppression)
- Built a front-end application to provide business users with the flexibility to customize campaigns each month
- Data Volume Considered: 100 GB of data per day; 100+ million records per day
Business Impact
- Identified USD 200,000 cost-benefit per month by bringing address standardization and in-house data processing
- Improved overall timeline for campaign preparation and execution by 95%
- Greater visibility and audit of data cleansing as compared to third party vendors
- The solution developed provided the flexibility to run any number of campaigns as opposed to only being able to run seven campaigns a month previously
