Chemical Conglomerate Transforms SAP Data in Snowflake for Analytic Insights
Customer's Challenge
One of the largest chemical conglomerates in the US was looking to modernize its analytics foundation by moving its SAP data from on‑premises SQL Server to Snowflake. They wanted to use this moment to establish an enterprise data model that would scale, reduce complexity, and improve trust as the business grew. They had the vision, just needed the expertise to get it done correctly.
phData's Solution
During a 10-week engagement, phData designed and implemented a modern modeling framework in Snowflake. Using a medallion architecture, with a Kimball‑style gold layer optimized for analytics, we implemented transformations in Coalesce while sharing best practices for Snowflake, Coalesce, version control, and deployment along the way.
The Full Story
A leading global provider of water and hygiene solutions was in the middle of modernizing its data platform from on-premises SQL Server to Snowflake. They had already begun landing SAP data in Snowflake, but it was raw and complex to use.
The legacy SQL Server models had “mostly worked”, yet they were inefficient, opaque, and error‑prone. Coalesce was the transformation tool of choice, but internal familiarity was limited. The future path to success was clear: build it right this time, but doing the work cleanly and transparently for the long term was where they needed help from phData.
We kicked off with a short, outcome‑driven plan.
In just 10 weeks, we designed the target models for the highest‑priority use cases, implemented them in Coalesce as a medallion architecture, and upskilled the client to sustain and extend the work.
To get there, we:
- Reviewed and analyzed the current state architecture, SAP data, and current data models.
- Designed the Snowflake information architecture utilizing a medallion (Bronze, Silver, Gold) architecture.
- Designed and implemented Kimball dimensional models in the Gold layer for the top 3 use cases.
- Stood up Coalesce and configured it with Snowflake instance, set up version control with GitHub, and created a deployment process
- Worked with the client teams to validate every aspect of the Gold layer
- Worked with the client’s analytics team to connect Tableau to the new Gold layer tables
- Delivered robust training sessions on both Coalesce and Snowflake
- Wrote documentation on best practices for both Coalesce and Snowflake
Why phData?
The chemical conglomerate chose phData because it didn’t just want an implementation of what it already had; it wanted a data model designed and done right. With phData’s extensive Snowflake, SAP, and Coalesce experience, the client knew they were in the right hands.
The chemical company also counted on phData to teach their internal teams how to use Snowflake and Coalesce efficiently so that they could eventually easily take over the entire process.
Results
Once the dimensional model for the highest-priority use cases was implemented in Coalesce, the client saw an immediate efficiency improvement over its previous SQL Server implementation.
Pipeline performance improved substantially, with SAP transformations moved to Snowflake and implemented through Coalesce using a medallion architecture with Kimball models. This delivered an estimated 40–60% faster execution and enabled earlier dashboard refreshes with tighter analytics SLAs for priority use cases.
This was complemented by full column-level lineage visibility in Coalesce, giving engineers and analysts the traceability needed to understand precisely how gold-level tables are constructed and to troubleshoot more quickly.
With the highest-priority work completed, the team leveraged reusable gold-layer patterns as blueprints for additional domains, reducing future data modeling and implementation timelines by 20–35% as they extended beyond the initial three use cases.
Reference Diagram
Take the next step
with phData.
Learn how phData can help solve your most challenging data analytics and machine learning problems.
