PSD Global BI Platform

The Challenge

A France-based international tyre dealer network needed visibility into purchasing volumes and overhead costs across 30 countries. Data from retail POS systems, e-commerce platforms, and intranet operations was scattered, making strategic and operational decision-making nearly impossible without manual consolidation.

The Approach

The project was delivered in four interconnected modules, each targeting a distinct data domain while sharing a common BI infrastructure.

• Tyre Manufacturer BI: Designed architecture and data model (logical/physical), built Pentaho Kettle ETL jobs to load dimension and fact tables, and created OLAP cubes with Schema-workbench for purchasing analytics.
• E-Commerce BI: Integrated daily transactional data from e-commerce platforms across 19 countries using Talend ETL, enabling cross-country sales comparison and trend reporting.
• Intranet BI: Automated collection and processing of intranet operational data into the centralised warehouse, providing HQ with an accurate worldwide purchase overview.
• Challenge BI: Extended the platform with competitive and promotional analytics, building PRD reports and scheduling them directly on the BI server.

The Results

International HQ gained a single platform to track purchasing volumes, overhead, and e-commerce performance across all 30 countries — enabling strategic projections previously impossible without manual data assembly.

Key Takeaways

Building shared OLAP schema and ETL conventions across four modules from the start saved significant rework. Consistent dimension design and naming conventions meant each new module could reuse existing structures rather than rebuild from scratch — compressing delivery time by over 30%.