Vesting Finance, a Dutch financial services organisation, was operating a complex on-premise data infrastructure that was expensive to maintain, difficult to scale, and increasingly unable to support the analytics and reporting requirements of a modern financial services business. Their data platform — built on legacy on-premise servers — required significant capital investment to maintain, delivered poor performance for complex analytics workloads, and lacked the governance and security capabilities required by their regulatory environment. Simultaneously, the organisation faced significant HR cost pressures and needed to improve the productivity of their data and analytics teams.
Vesting Finance's data infrastructure had been built incrementally over more than a decade, resulting in a complex, fragmented on-premise environment that was expensive to operate and difficult to evolve. The infrastructure comprised multiple on-premise servers running a combination of SQL Server databases, legacy ETL tools, and a collection of Excel-based reporting processes that had accumulated over years of tactical problem-solving.
The operational challenges were significant. Data pipeline failures were frequent and difficult to diagnose. Analytics workloads that should have completed in minutes were taking hours. The data team spent the majority of their time on infrastructure maintenance and firefighting rather than delivering analytical value. And the governance and security capabilities of the on-premise environment were insufficient for the regulatory requirements of a financial services organisation.
The financial case for migration was clear: the on-premise infrastructure required significant capital investment to maintain, and the operational costs — including the staff time required to manage it — were substantial. But the migration needed to deliver more than cost savings. It needed to deliver a genuinely better data platform — one that could support the organisation's analytics ambitions and meet its regulatory obligations.
The HR dimension added further complexity. The existing data team had skills and experience calibrated to the on-premise environment — SQL Server administration, legacy ETL tools, and manual reporting processes. The new cloud environment would require different skills, and the organisation needed a plan for managing the transition.
MDN.digital was engaged to design and execute the migration, covering architecture design, implementation, data migration, and the development of a strategic personnel plan.
The target architecture was designed around the Azure Lakehouse pattern — combining Azure Data Lake Storage Gen2 for scalable, cost-effective storage with Databricks for compute, transformation, and analytics. Microsoft Purview was selected as the governance layer, providing automated data classification, lineage tracking, and a unified catalogue of all data assets. Power BI was selected as the analytics and reporting platform, replacing the legacy Excel-based reporting processes.
The medallion architecture was adopted for organising data in the Lakehouse: Bronze for raw source data, Silver for cleaned and validated data, and Gold for business-ready aggregations. This structure provided clear separation of concerns, simplified governance, and made the data platform understandable to both technical and business users.
The architecture was designed with cost optimisation as a primary constraint. Every architectural decision was evaluated against its cost implications, and the design incorporated the cost optimisation patterns that would deliver the target 50% cost reduction.
The first phase established the Azure infrastructure and governance foundation. Azure Data Lake Storage Gen2 was provisioned with appropriate security controls — private endpoints, encryption at rest, and role-based access control. Microsoft Purview was configured to scan and classify all data assets, providing immediate visibility into the personal data holdings that would need to be governed under GDPR.
Databricks Unity Catalog was configured as the governance layer for all Databricks-managed assets, providing fine-grained access control, automated lineage tracking, and a searchable data catalogue. The integration between Unity Catalog and Microsoft Purview provided a unified governance view across the entire Azure data estate.
Azure DevOps was configured for CI/CD of data pipelines, establishing the engineering practices required for reliable, reproducible data platform operations.
The data migration was executed in waves, prioritising the data domains with the highest business value and the most significant performance problems. Each wave involved:
The migration of legacy ETL processes to Databricks Delta Live Tables delivered immediate performance improvements. Pipelines that had taken 4–6 hours to complete on the on-premise infrastructure completed in 45–90 minutes on Databricks — a 4x performance improvement that freed significant capacity for analytical work.
The analytics modernisation phase replaced the legacy Excel-based reporting processes with Power BI dashboards connected to the Gold layer of the Databricks Lakehouse. This involved:
The transition to Power BI self-service analytics was transformative for the organisation. Business users who had previously relied on the data team for every report could now create their own analyses, significantly reducing the demand on the data team and improving the speed of business decision-making.
Alongside the technical migration, MDN.digital developed a strategic personnel plan to address the skills transition and HR cost reduction objectives.
The plan was based on a detailed skills assessment of the existing data team, mapping current skills against the requirements of the new cloud environment. The assessment identified three categories of team members: those whose skills were directly transferable to the new environment (typically data analysts and business intelligence developers), those who could be upskilled to the new environment with targeted training (typically data engineers with SQL and ETL experience), and those whose skills were primarily calibrated to the legacy on-premise environment with limited transferability.
The personnel plan addressed each category differently. For transferable skills, the plan focused on targeted upskilling — Azure certifications, Databricks training, and Power BI development skills. For upskillable team members, the plan defined a structured development programme with clear milestones and assessment criteria. For team members with limited transferability, the plan defined a managed transition process — including redeployment to other parts of the organisation where their skills were relevant, and where redeployment was not possible, a structured offboarding process.
The result was a smaller, more capable team — with higher average productivity, stronger cloud and analytics skills, and a lower total HR cost. The 30% HR cost reduction was achieved through a combination of team size reduction and the elimination of contractor and vendor costs associated with maintaining the legacy infrastructure.
The migration delivered against all three primary objectives: cost reduction, capability improvement, and HR optimisation.
The 50% reduction in infrastructure costs was achieved through a combination of architectural optimisation (right-sized compute, storage tiering, spot instances) and the elimination of the capital and operational costs of the on-premise infrastructure. The ongoing cloud costs are significantly lower than the equivalent on-premise costs, and the cost optimisation practices implemented during the migration provide a foundation for continued cost discipline as the platform grows.
The capability improvements were equally significant. The data team's time allocation shifted dramatically — from 70% infrastructure maintenance and 30% analytical work before the migration, to 20% infrastructure operations and 80% analytical work after. The self-service analytics capability delivered to business users has accelerated decision-making and reduced the backlog of analytical requests.
The governance improvements delivered by Microsoft Purview and Unity Catalog have strengthened Vesting Finance's regulatory compliance posture. The automated data classification and lineage tracking provide the visibility into personal data holdings required for GDPR compliance, and the access control framework ensures that sensitive financial data is accessible only to authorised users.
The strategic personnel plan delivered the targeted HR cost reduction while maintaining — and in most areas improving — the analytical capability of the team. The upskilling programme has created a team with strong cloud and modern data platform skills, reducing the organisation's dependence on external consultants for ongoing platform operations.
MDN.digital designed and executed a comprehensive migration from on-premise to Azure cloud, including the implementation of a Databricks Lakehouse architecture, Microsoft Purview for governance, and Power BI for analytics. Alongside the technical migration, MDN.digital developed a strategic personnel plan to improve team productivity while reducing HR costs.
Majid Al Futtaim, one of the Middle East's leading retail and real estate conglomerates, sought to transform their custo...
Following a significant data security incident in 2026, Odido — one of the Netherlands' leading telecommunications provi...