A Business Intelligence platform that consolidates the sales and purchases of four companies into a single model, refreshed every day.

We built a relational Data Warehouse with a star schema that places the four companies under one analytical language: sales and purchase facts (quantity, amount, cost, margin) and customer, product, supplier and province dimensions. What could not be compared before is now read in the same table.

An ETL process reads each source in its own format, whether an Excel file, a SharePoint folder or a direct ERP connection, and normalises it into the model. It runs every night without intervention. When management opens the report, the previous day's data is already available.

We replaced files and shared folders with a permanent SQL connection to the source databases over a private network. We studied each database, configured the VPN and left a reliable, daily refresh in place.

The reports analyse performance by customer, area, team, salesperson and product, with geographic and time analysis. They include margin, active customers, moving averages and year-on-year comparisons, plus fiscal-year filters. The same architecture serves the purchases report.

The group's management committee reviews the sales and purchases of all four companies in the same report. The manual consolidations that delayed every close are gone.

The automatic process refreshes the information every night. Nobody builds reports by hand and transcription errors have been eliminated.

Customer, area, product, salesperson and margin are analysed with the same criteria in every company. Questions that once took days are answered instantly.

Adding a company, an ERP or a new analysis domain means extending the model, not starting over. The infrastructure runs with daily backups and without the client having to manage the technical layer.