Databricks offers a Rolls-Royce experience for data teams. It’s powerful, flexible, and almost limitless in possibilities. But with great power comes great responsibility: costs can spiral quickly if you don’t have the right strategy. 

So, how do you optimise your spend, avoid costly mistakes, and ensure your Databricks investment delivers maximum ROI? 

This was the focus of SGI’s first Databricks event in Belgium, held on 21 October in Brussels, where leading experts shared actionable insights. 

Meet the Experts 

We were joined by two incredible speakers: 

Why Data Quality Is Back in the Spotlight 

A decade ago, businesses focused on BI, reporting, and analytics, with data management and quality as key priorities.  

Those who ignored these fundamentals are now struggling to make sense of LLMs and AI. Once again, data quality and management are at the forefront, as they are critical for change management and any successful data strategy. 

Another surprising insight? Many companies have 8, 9, or even 10 separate data strategies. It sounds crazy, but it happens often. 

The solution is simple: 

Pick one strategy. Do it right. Do it well. Any decision is better than indecision, especially when it comes to data. Choose a clear path, align everyone on the why and outcomes, and work toward a defined end goal. 

Technical Strategies for Cost Optimisation

Axel shared practical tips based on real-world experience: 

1. Increase Data Literacy 

Understand your organisation’s data maturity. Axel outlined five levels: 

Belgian businesses often aim to be “data-driven”, but the journey requires awareness and education. 

2. Use the Medallion Architecture 

Databricks’ Medallion architecture improves data quality by structuring pipelines into three layers: 

Bronze (raw data) → Silver (validated data) → Gold (enriched data) 

This design pattern ensures data is progressively cleaned, transformed, and prepared for analysis, creating a single source of truth for reporting, predictive analytics, and machine learning. 

3. Avoid Costly Replication 

During his presentation, Axel highlighted that data source replication is usually to blame for the increase in costs. To prevent this additional cost, he suggested one of 2 ways to approach this:   

4. Monitoring & Alerting 

5. Axel’s Quick Cost Checklist 

Business Perspective: How to Calculate ROI of Technology 

Sam Rédelé, our second speaker, delivered an insightful presentation on business strategies for data leaders, sparking lively discussion during the networking session.  

He showed that technology ROI should be driven by business enablement, not only by tool analysis. To do this, your tech ecosystem must first mirror how the business really works, which means connecting systems and breaking down silos. 

Sam introduced knowledge modelling on top of the existing data ecosystems. This is a business layer that connects systems that previously couldn’t talk to each other, unlocking faster interoperability and enabling systemic, end-to-end analytics rather than siloed reporting. 

This knowledge-driven foundation enables an AI-First operating model with real-time operational improvements and smart decision-making.