Wein & Vinos The road to a modern Azure Databricks architecture

The considerable growth of e-commerce means that IT infrastructures which have evolved over time inevitably reach their scalability limits. When these technological constraints are then exacerbated by a fixed group-wide decommissioning schedule, the IT personnel responsible can find themselves with few options.

This was precisely the challenge faced by Wein & Vinos GmbH, Germany’s leading specialist supplier of Spanish wines and part of the Hawesko Group. The main task: the complete removal of the company’s SAP BW legacy system and the parallel setup of a future-proof data lakehouse within just seven months.

The Challenge: Data growth collides with hard-and-fast infrastructure deadlines

What began as a small wine shop in Berlin’s Kreuzberg district has grown into a leading e-commerce specialist with over 1,600 wines in its assortment and around 180 employees. The incredible scalability of Wein & Vinos’ business model was demonstrated most recently during the Covid pandemic: the company’s turnover increased by almost 30 percent to nearly 62 million euros, with up to 6,000 packages leaving its warehouse on peak days.

This scalability was reflected directly in the data volumes handled by the company’s previous systems. The local Microsoft Dynamics NAV system included core tables with, at times, more than 219 million data records. HubSpot delivered high-volume CRM data, and Google Analytics recorded over 150 million session entries. SAP BW served as Wein & Vinos’ central data warehouse for bundling these data streams.

The need for action with regard to technology became acute, however, when a hard-and-fast decision was made at group level: the Hawesko Group decided to completely decommission the group’s SAP BW system by 31 May 2026 at the latest. This gave Vinos an immovable migration window from November 2025 to May 2026.

The Cause: The reverse-engineering dilemma in legacy systems

Previously, traditional data warehouse structures were sufficient for evaluating structured business data in isolation. Today’s dynamic e-commerce processes, however, require the flexible and timely integration of highly heterogeneous data sources.

The biggest technical hurdle for this migration project, however, was not only the sheer volume of data involved. The existing processes, which had evolved over time, were also extremely complex, and a total of 56 loading threads had to be migrated. In order to do this, the old SAP BW logic – including 17 highly complex ABAP and SQL routines deeply embedded in the old system landscape – had to be flawlessly reverse engineered.

The Solution: Setup of a modern data lakehouse architecture based on Azure Databricks

In order to position themselves for future success and to optimally leverage existing synergies within the Hawesko Group, Wein & Vinos decided to switch to a state-of-the-art data lakehouse architecture. This approach enabled them to combine the strict governance structures and structured data models of a traditional data warehouse with the tremendous scalability and flexibility of a data lake.

To handle the complex implementation and ensure a frictionless migration, Wein & Vinos enlisted the help of inovex, an experienced IT partner for digital transformation projects.

The project was based on a state-of-the-art technology stack. Azure Databricks and Azure Data Lake Storage were used to create a storage and processing infrastructure to handle the enormous quantities of data, thus ensuring a high-performance data basis. The structured transformation and versioning, as well as the seamless documentation of the data models, were implemented via data build tool (dbt).

In order to ensure agile and transparent project development supported end-to-end by CI/CD, the entire process was managed via Azure DevOps.

The prepared data was ultimately transferred to Power BI, which served as a centralised front end for business reporting and enabled well-founded, data-based operations management.

Fit for the Future

The successfully completed project enabled Wein & Vinos to do much more than simply meet the system migration deadline. The company also laid the technological foundation for future innovations. The new data lakehouse architecture eliminated former bottlenecks and created the ideal basis for establishing forward-looking use cases involving advanced analytics and machine-learning. From recommending the perfect wine to a customer to setting dynamic pricing or accurately predicting product demand, Wein & Vinos has managed to turn its infrastructure issues into a strategic competitive advantage. The company is now optimally equipped for a data-driven e-commerce future.