ON-POINT Insight #09 - expertise from the consulting frontline
Snowflake + Data Vault — Building a Scalable Data Platform by Dharshan Seesurn
In a recent project, I helped design and deliver a scalable cloud-based data platform using Snowflake at its core. The goal? To unify semi-structured data from multiple business units into one system — powering reporting, analytics, and strategic decision-making.
Automated ELT at the Core
At the heart of the platform was an automated ELT pipeline pulling data from multiple source systems. Built with modern tooling and principles, it was designed to be fast, repeatable, and hands-off.
Data landed in Snowflake, was transformed using dbt and flowed into a Data Vault architecture.
Why Data Vault?
Data Vault provided the structure we needed to combine flexibility with governance:
Performance at scale — optimised Snowflake virtual warehouses for concurrency and cost-efficiency
Incremental loads — reduced compute time while keeping data fresh
Data marts — designed for fast retrieval tailored to business users
Materialised views — ensured consistently high query performance
We also needed a clear, auditable separation between:
Raw data (Raw Vault)
Business logic (Business Vault)
Reporting views (Data Marts)
Using dbt meant we could create modular, version-controlled SQL models while managing dependencies between transformation layers. The result: faster development, easier debugging, and more reliable deployments.
Start with Stakeholders, End with Insights
The journey began with stakeholder workshops to align on reporting needs and KPIs. From there, data was staged, modelled, validated and transformed into business-friendly structures. BI dashboards were tailored to user roles and designed to answer critical business questions.
From Data Vault to Business Insights
To meet BI and dashboarding needs, I designed a star schema in the data mart layer — clean fact and dimension tables optimised for performance, ease of use and consistent reporting across tools.
Solving for Complexity and Compliance
Two big challenges shaped our approach:
- Full historical traceability — handled via satellites in the Data Vault model, capturing every change with timestamps, metadata and full audit trails.
2. Performance at scale — maintained through:
Optimising Snowflake virtual warehouses for concurrency and cost-efficiency
Using incremental loads to reduce compute time
Designing data marts for fast retrieval
Applying materialised views for high-performance queries
From Complexity to Clarity
By combining Snowflake’s scalability with the structure of Data Vault, we built more than just a data platform — we created a foundation for trust, compliance, and insight. The result is a system that not only meets today’s reporting and regulatory needs but is ready to adapt as business demands evolve.
At ON-POINT, we bring consulting expertise from the front line — helping you deliver complex projects with confidence. Reach out to Frank Lehmann or Tobias Reuter to start the conversation.