North American CPG & Retail Giant Centralizes Business Unit Data to Analyze Direct Spending
Customer's Challenge
A giant North American CPG & Retail company sought to centralize and analyze its direct spend data across disparate business units.Â
phData's Solution
phData implemented a custom data vault solution to centralize data from the first business unit’s SAP ECC system into the Snowflake Data Cloud. This setup, supported by custom integration code, enables continuous analysis of overall and unit-specific direct spending.Â
Results
phData successfully deployed the data vault and the ingestion code and migrated the initial business unit. Future business units are now in progress for migrating to the new platform, allowing for effective direct spend data across multiple business units.Â
The Full Story
A large CPG & Retail conglomerate, with business units centered on the home, security, and commercial building markets, was unable to analyze direct spend data across its disparate business units due to disjointed data warehousing systems and processes.
The customer approached phData to partner on a solution, and together, the teams established three objectives for a pilot project focused on one of the company’s business units.
- Objective 1: Establish a foundation using data vault, DataOps (CI/CD), and dbt.
- Objective 2: WINN direct spending data from SAP data modeled using data vault and loaded into Snowflake.
- Objective 3: Knowledge transfer to cross-training for the internal customer team.
Why phData?
The customer needed a partner with experience in data engineering, strategy, and migration to provide guidance and technical assistance. The client chose phData due to our reputation for fast-tracking customer projects while providing the necessary tools to ensure long-term success.
Streamlining Data Management and Reporting
To accomplish the first objective, phData worked alongside the customer to codify and control the business logic, enforce the business rules on files, control updates via version control & continuous integration, and improve data quality. Â
For objective two, phData and the customer loaded the live data from SAP ERP into Snowflake using a productionalized SAP extractor. This allowed multiple systems to load quickly.Â
Business logic, previously not documented or stored in various locations, was codified and controlled in one place, then enforced on the files upon loading, which noticeably improved data quality.Â
Simultaneously, a data model was structured to accommodate changes, which resulted in a 99% data match. Leveraging Tableau and Snowflake, the customer could then produce accurate direct spending reports.Â
In the project’s final phase, objective three was met through a series of knowledge transfer sessions between the customer and phData that included an architectural overview, a runbook with steps to take when pipelines fail, and a step-by-step product development guide.Â
Moving forward, the client now has the necessary tools and infrastructure to build the remaining business unit data — all in one centralized location for the entire business to produce accurate and comprehensive direct spending reports.
Take the next step
with phData.
Learn how phData can help solve your most challenging data analytics and machine learning problems.
