Building the data foundation for a European food distributor
The distributor was running their business on revenue numbers and gut feel. ERP, CRM, and forecast data lived in separate systems, and questions about margin, mix, or forecast accuracy took a week of manual work to answer.
The challenge
Leadership wanted to be data-driven but had no infrastructure to back it up. ERP, CRM, and forecast data lived in three separate systems with no clean way to combine them. Margin by customer, product mix shifts, and forecast accuracy were all questions the team simply couldn't answer; the data existed, but it was scattered.
Our approach
We built a Microsoft Fabric lakehouse with incremental ingestion pipelines from their ERP, CRM, and Excel forecast files. PySpark notebooks handle the ETL into a medallion architecture, bronze, silver, gold, with a clean semantic layer on top. Power BI sits over the gold layer, giving the team self-serve access to combined data for the first time.
What we delivered
- Microsoft Fabric lakehouse with bronze/silver/gold medallion architecture
- Incremental ingestion pipelines from ERP, CRM, and Excel forecast files
- PySpark notebooks for ETL, fully documented for the internal team
- Power BI semantic layer and reports giving leadership full P&L visibility
The outcome
Before, the team was effectively blind to questions leadership wanted to answer: cost-to-serve by customer, margin shifts inside product lines, where forecast accuracy actually broke down. Not because the answers took too long. The data existed, but it lived in three disconnected systems. Now they can see it. One trusted source of truth, refreshed within seconds, with the architecture in place to add predictive models on top.
“Our finance team spent the first week of every month chasing operations for the right margin numbers. Conlatio rebuilt the data foundation so margin by customer and product line is on the dashboard before the books close. The next phase, alerts when margin slips on a specific line, runs on the same architecture we built. I finally trust the numbers I share with our leadership.”
More case studies
Back to all case studies
BI infrastructure and AI-driven order automation for a German food and beverage producer
The producer needed business visibility and relief from time-consuming manual order intake. Incoming orders arrived as faxed and emailed PDFs, all keyed in by hand into the ERP.
Read the case study →
Securing contracts by eliminating manual reporting for an insurance broker
A growing broker was operating on fragile Excel documents and Python scripts, sending unreliable data to clients and risking crucial contracts.
Read the case study →