Jaco van der Laan

From Whiteboards to Working Models: How We Built a Model-Driven Data Warehouse at de Volksbank with…

19/09/2025  ·  Data Architecture  ·  Data Engineering  ·  Data Modeling

article base/metadata-models
‹ ›

from-whiteboards-to-working-models-how-we-built-a-model-driven-data-warehouse-at-de-volksbank-with

Why teamwork and model-driven engineering delivered a future-proof data warehouse under pressure

Summary

At de Volksbank/ASN, we faced the challenge of replacing the Risk & Finance Data Warehouse under strict ECB/DNB requirements. Working with CrossBreeze, we applied Model Driven Data Engineering (MDDE) with PowerDesigner data models as the foundation.

Alongside PowerDesigner, we relied on CrossBreeze’s CrossTest for automated validation and the open-source PowerDecomposer to unlock metadata-driven automation.

The lessons we learned also contributed to the development of CrossModel and CrossGenerate as next-generation tools.

This case study reflects on what we accomplished, why teamwork was essential, and why MDDE is truly a team sport.

The Challenge

Replacing a Risk & Finance DWH is not business as usual. The demands were steep:

  • Meet regulatory deadlines from ECB/DNB.
  • Provide lineage and governance that auditors could trust.
  • Ensure quality and consistency across multiple domains.
  • Deliver a platform that was faster and more maintainable than the old ETL-heavy system.

Traditional approaches — hand-coded ETL, scattered documentation, manual testing — simply wouldn’t scale.

The Model Driven Approach

Instead of coding everything by hand, we adopted Model Driven Data Engineering (MDDE):

  • PowerDesigner models became the single source of truth for structure, mappings, and business rules.
  • Metadata repository captured everything — no hidden logic.
  • Generators and automation frameworks turned models into SQL code and automated tests.
  • CI/CD pipelines ensured deployments were reproducible and safe.

The idea was simple but powerful: if it’s in the model, it’s in the code.

The Role of CrossBreeze

In our project we heavily used CrossTest, CrossBreeze’s testing framework. It allowed us to embed automated regression and functional tests directly into the data warehouse pipelines, giving us confidence in every release.

We also relied on PowerDecomposer, an open-source solution created by CrossBreeze: 👉 PowerDecomposer
This tool decomposed PowerDesigner models into usable metadata structures, making it possible to connect modeling with automation and governance. It filled an important gap and was essential for making PowerDesigner fit into our model-driven approach.

At the same time, CrossBreeze was already developing CrossModel and CrossGenerate — new tools born from the very challenges we faced.

  • PowerDesigner, while a great product, had its limits: SAP no longer prioritized it, development had slowed, and gaps were evident in metadata management and automation.
  • Based on insights from our Volksbank project (and other customers), CrossBreeze began developing CrossModel as a modern replacement for PowerDesigner, Erwin, and other legacy modeling tools.
  • Alongside it, they started building CrossGenerate to generate code directly from metadata models.

In this way, the Volksbank project was not only a success in itself but also part of the evolution of new model-driven engineering tools.

Results That Spoke for Themselves

The combination of bright people, solid tools, and disciplined modeling delivered remarkable results:

  • 3x faster development speed compared to traditional coding.
  • Onboarding time dropped from 3 months to just 3 weeks.
  • 10,000+ automated tests and controls embedded into the platform.
  • Production go-live in 2.5 years, meeting all regulatory deadlines.

Most importantly, we had confidence: in the data, in the lineage, and in the reproducibility of every release.

Why the Team Was the Key

Looking back, the technology was essential — but the real success came from the team dynamic:

  • Volksbank colleagues who believed in modeling-first.
  • CrossBreeze experts who brought deep product knowledge.
  • A shared vision: move from whiteboards to working models.

Complex data solutions demand collaboration. No single person could have solved this puzzle. It took a team of bright, committed people, working toward the same goal.

Lessons Learned

For me, this project was proof of three things:

  1. Model-driven engineering works — when models drive everything, quality and speed improve.
  2. Automation pays off — code generation and testing aren’t luxuries, they’re necessities.
  3. Teamwork is the multiplier — bright people, aligned around the same vision, make the difference between failure and success.