...

get in touch /

helping enterprises become AI-native organizations

Data Platform Modernization for a Major European Frozen Food Delivery Company

The client is one of Europe’s leading direct-to-consumer frozen food delivery companies, operating 150+ branches and servicing customers across multiple European countries. With over 9,000 point-of-sale terminals generating continuous transactional data, the business relies on fast, accurate reporting to manage daily operations across its distributed network. The company’s data landscape includes CRM, PIM, Workday, order management, and tour planning systems — all of which needed to work in concert.

/Key Results

  • 82% cost reduction: Compute cost dropped from €43/h to €7.74/h
  • 67% faster BI reports: Power BI full refresh reduced from 30 min to 10 min
  • 40% faster data load: End-to-end data load time cut from 14 min to 8.5 min
  • New capability: Incremental Power BI refresh introduced at 40–50 seconds (previously unavailable)
  • GDPR compliance: Full PII deletion mechanism and audit trail — regulatory risk eliminated
  • Reliable data SLA: Data available by 6:30 AM daily, every day — replacing an unpredictable schedule

/The Challenge

The client’s Oracle Exadata data warehouse had reached end-of-life, and the cost of maintaining it on Oracle/Synapse — at €43 per hour with no cloud elasticity — was unsustainable. Every morning, branch managers across Europe were unable to make informed decisions until Power BI reports finished their 30-minute refresh cycle. Meanwhile, the absence of a PII deletion mechanism left the business exposed to GDPR liability, and the fragmented system landscape made any future modernisation effort increasingly complex.

  • High compute cost: €43/h with no ability to scale down during off-peak hours
  • BI bottleneck: 30-minute full refresh blocking morning operations across 150+ branches
  • GDPR exposure: No mechanism for PII deletion or audit trail
  • Siloed systems: CRM, PIM, Workday, Order Management, and Tour Planning operating independently
  • Scale constraints: 9,000+ POS terminals generating continuous data with no unified ingestion layer

/Project Goals

  • Migrate off Oracle Exadata: Move to a modern, cloud-native data platform before end-of-life
  • Reduce operating costs: Achieve meaningful compute cost savings through cloud elasticity
  • Accelerate BI reporting: Deliver Power BI reports fast enough to support morning operational decisions
  • Achieve GDPR compliance: Implement PII deletion and a full audit trail
  • Unify the data landscape: Create a single ingestion framework for all source systems
  • Enable AI/ML readiness: Build a foundation capable of supporting future analytics and ML workloads
  • Zero business disruption: Complete the migration without interrupting daily branch operations

/mindit.io’s Solution

mindit.io architected and delivered a full end-to-end migration from Oracle Exadata to a Databricks Lakehouse on Azure. The solution was built around a medallion architecture (Bronze, Silver, Gold) that enforces data quality at every tier, and a metadata-driven ingestion framework that makes it straightforward to add new source systems as the business grows.

To eliminate business risk during the transition, the team ran the old and new platforms in parallel throughout — meaning a fallback was always available and daily operations were never disrupted. The Sales domain was migrated first, delivering measurable business value early and validating the architecture before broader rollout.

  • Medallion Architecture: Delta Lake with Bronze / Silver / Gold tiers and data quality checks at every layer
  • Metadata-Driven Ingestion: Extensible framework with built-in validation — new sources added with minimal effort
  • 29 ETL processes migrated: Covering raw vault, business vault, and mart central layers
  • Power BI Optimisation: Full semantic layer rebuild with incremental refresh reduced to 40–50 seconds
  • GDPR Compliance Layer: Automated PII deletion with complete audit trail
  • Parallel running: Legacy and new platform operated simultaneously throughout the migration
  • Full knowledge transfer: Complete documentation delivered; client team fully self-sufficient at handover

Technologies used: Databricks · Delta Lake · Photon · Unity Catalog · Azure Data Lake Gen2 · Azure Data Factory · Power BI · Azure DevOps · Databricks Asset Bundles · CI/CD · Oracle Exadata · Salesforce · SAP · Workday

/Results

  • Compute cost: €43.48/h → €7.74/h (82% reduction)
  • Power BI full refresh: 30 min → 10 min (67% faster)
  • Raw vault load: 7 min 11 s → 4 min 50 s (33% faster)
  • Business vault load: 7 min 6 s → 3 min 47 s (47% faster)
  • Incremental refresh: Not available → 40–50 sec (new capability)
  • Data SLA: Unpredictable → 6:30 AM daily (reliable)
  • GDPR status: Non-compliant → Full compliance + audit trail

/Customer Testimonial

“The new architecture is ready for AI/ML workloads and Databricks Apps — turning the platform from a cost centre into a foundation for future growth.”

/Strategic Takeaways

  • Modernisation is a sequencing problem: Running old and new platforms in parallel — rather than cutting over all at once — is what made zero-disruption possible. Most migration failures come from big-bang switchovers.
  • Domain-first delivery accelerates value: Migrating the Sales domain first gave the business tangible results within weeks, not months — building confidence and validating the architecture before full rollout.
  • Metadata-driven frameworks pay dividends over time: Building extensibility into the ingestion layer from day one means new data sources can be onboarded in days, not sprints. This is infrastructure that compounds.
  • Governance and compliance are not afterthoughts: Embedding GDPR controls directly into the platform architecture — rather than bolting them on later — eliminated regulatory risk permanently and reduced future compliance overhead.
  • Cloud elasticity changes the cost model fundamentally: An 82% reduction in compute cost was only possible because the new architecture scales to zero when demand drops. Legacy on-premise platforms simply cannot do this.

/What’s Next?

Whether you’re running a legacy data warehouse that’s reached end-of-life, facing GDPR exposure, or struggling with BI reports that can’t keep up with your operations — this is exactly the kind of challenge mindit.io is built for.

Ready to explore what’s possible for your organisation? Get in touch with our team →

  • / foodtech

A major European frozen food delivery company faced a critical inflection point: its legacy Oracle Exadata data warehouse was nearing end-of-life, Power BI reports were taking 30 minutes to refresh every morning, and GDPR compliance gaps were creating regulatory exposure across 150+ branches. mindit.io designed and executed a full migration to a Databricks Lakehouse on Azure — reducing compute costs by 82%, cutting BI refresh times by 67%, and delivering a GDPR-compliant, AI-ready platform that now serves 9,000+ POS terminals reliably by 6:30 AM every day.

/ turn your vision into reality

The best way to start a long-term collaboration is with a Pilot project. Let’s talk.

Request a Pilot
Get in Touch