Data Platform Standardization for a Diversified Company

The situation

A diversified company with operations across multiple business units had standardized on Databricks as its data platform but had done so unevenly. Regional deployments had diverged in cluster configuration, job design, and governance practices. The data team was capable but stretched thin, and there was no consistent architectural baseline across the platform. At the same time, the company had made a significant investment in an Oracle ERP that had not delivered the reliability the business expected, and internal confidence in the system was low.

The data team needed outside perspective on two things: whether the Databricks platform could be stabilized and standardized without a rebuild, and how to manage analytical work that sat adjacent to a troubled ERP without inheriting its problems.

What we built

CorrDyn started with a data assessment of the Databricks environment, reviewing cluster configurations, job scheduling patterns, cost allocation, and governance across regional deployments. The assessment produced a prioritized set of recommendations covering architectural inconsistencies and cost optimization opportunities, and became the blueprint for the standardization work that followed.

From January 2024, CorrDyn provided staff augmentation to support a global initiative to standardize the Databricks platform across regions. The engagement operated on an annual budget for ongoing support and decision-making, with separate capital budget for larger project work. Within that structure, we completed a customer segmentation project: a model built in Databricks that assigns segment labels to customers based on attributes and transaction patterns, then pushes those labels back into Salesforce as custom fields. Sales and marketing teams can now filter and target by segment without leaving their CRM, and the model refreshes on a schedule to stay current.

On the ERP side, CorrDyn provided advisory input as the company worked through Oracle adoption challenges. The consistent recommendation was to minimize customization and build custom logic in the data layer outside the ERP, where it can be maintained independently of vendor release cycles.

What changed

The Databricks platform moved from a collection of inconsistent regional deployments to a standardized architecture with shared governance practices. Customer segmentation data reached the sales and marketing teams in Salesforce for the first time, replacing manual segmentation done in spreadsheets. The ERP advisory gave the data team a clear position on where to draw the line between platform configuration and custom development. The engagement has continued past its initial scope as a fixed-budget advisory relationship, with CorrDyn supporting ongoing decisions as the platform grows.