Our client, a holding company with four operating divisions, generated annual revenue of over €100 million and employed around 150 people per division. Over the years, each division had built its own ERP system — SAP here, Microsoft Dynamics there, supplemented by custom-built solutions and countless spreadsheets. The result: an impenetrable thicket of data silos that made consolidated reporting at holding level nearly impossible.
The data hoarding problem
The core issue was not technical but structural: knowledge was being hoarded. Not out of bad intent, but because no shared data infrastructure had ever been built. Each system spoke its own language, each division its own controlling vocabulary. A CFO wanting to see a cross-divisional margin had to wait weeks for manual reports — or build them manually themselves.
Does this sound familiar?
If your data sits locked in a legacy system and every analysis starts with an export, let's look together in a free first conversation at how to open up that access.
First step: Tableau as a shared visualization layer
As a first step, we introduced Tableau as a unified BI layer. The idea was not to immediately replace all systems, but to make them readable. Tableau acted as a translator between the different ERP environments and gave management a unified dashboard for the first time — showing revenue, margin, and resource use across all divisions, live, without manual steps in between.
Java connector: the technical bridge
The real challenge was data collection. The legacy ERP systems offered no modern API interfaces. We developed a Java-based connector that accessed the databases of the various systems directly, extracted and transformed the raw data, and loaded it into a central Snowflake data warehouse. From there, Tableau drew its dashboards in real time.
The results spoke for themselves: the ERP was connected, historical data migrated, and the previously isolated time-tracking system integrated into the central data stream. What had previously taken weeks was now an automated daily close — and the CFO could finally see at the press of a button what was actually happening in the company.
Knowledge locked inside systems isn't knowledge — it's dead weight. The first step toward data-driven management is setting it free.