Case Studies | Analytics, BI, Data Integration, Cloud Services, ERP

Case Study: Kimberly-Clark - Azure Synapse Analytics on POS

Written by KPI Partners News Team | Apr 12, 2021 1:31:42 PM

Leveraging Azure Synapse Analytics for Point-of-Sale data analysis and reporting

About Kimberly-Clark

Kimberly-Clark is an American multinational personal care corporation with 2020 revenue of over $19 billion.

Customer Need / Business Driver

Kimberly-Clark’s Commercial Sales and Retail Sales organizations struggled to integrate point of sale (POS) data sets from their channel business partners. Also, because the data was scattered across many systems, Kimberly-Clark users had no single place to go for their analytic needs.

Selection Process

Kimberly-Clark selected KPI over multiple vendors through a rigorous RFP process. KPI’s expertise in Hadoop, Azure, Power BI, and data and analytics in general was the key differentiator for Kimberly-Clark. Also, KPI’s blended shore model minimized cost and risk for Kimberly-Clark.

What KPI Delivered

An Azure based solution for the POS data coming from disparate source systems. This system reduced the ETL run time by over 80% from Kimberly-Clark’s prior system. It also reduced report execution time by over 90%.

Because Azure Synapse Analytics is a complete solution for blending data and reporting, it served as the perfect solution for the different personas across the organization. Because Synapse Studio was a new service, combining Spark and data warehousing capabilities, KPI delivered best practices and a re-usable architecture. KPI also held knowledge transfer workshops.

PoC Architecure

Business Benefits

  • A collaborative platform for different personas
  • Optimized process for reporting Point of Sale data
  • Data Standardization
  • Faster data processing with Azure Data Factory and Synapse Spark, which cut data load times by over 80%
  • A scalable, reusable, modern architecture
  • A universal format that allowed for easy integration with downstream systems, access by data scientists, as well as access by traditional data warehousing users. This format also supported incremental loads and slowly changing dimensions
  • A simple mechanism for reporting both raw and data warehousing data