A global medical technology company had invested in Azure Databricks and Power BI as the foundation for their analytics capability. The right platforms were in place but there were concerns about the ROI. Efforts were ineffectively focussed on a like-for-like migration from Qlik, business adoption was low, and the data team had become a bottleneck.
Analytics8 was already working with the organisation on separate initiatives, which gave us the operational context and relationships required to support on this issue. We conducted a focused assessment with structured conversations with the Head of Data, analysis of the platform setup, and an opinionated view of what was working and what wasn’t.
The Aim
The organisation had made solid progress with Databricks and Power BI. Unity Catalog was enabled. A CI/CD deployment process was in place. Initial dashboards had reached production. But the expected value wasn’t arriving at the expected pace – and without a clear benchmark, it was difficult to know whether these were teething problems or something more structural. Previous consultancy support had not been able to unlock the value. They needed an honest, trusted, and experienced view of where things stood and a clear sense of what to address first.
Challenges
Our assessment identified challenges across four dimensions; people, process, technology and data.
- People and adoption: The Databricks and Power BI initiatives had been positioned primarily as IT projects. Business teams had limited platform access, insufficient training and no visibility into what data was available. With no communities of practice and no internal champions, nothing was pulling adoption forward. The data team was carrying everything and paying the price in velocity.
- Process and ways of working: The CI/CD pipeline and release process were creating UAT bottlenecks. There was no structured change management, meaning developers could alter data models and reports without communicating downstream impact. The catalog structure lacked domain separation and sandbox environments to support safe development.
- Technology and data model: Most data consumed by Power BI was served via views – often views built on views – with computation happening at query time. The result was slow report refreshes that compared poorly to Qlik, undermining confidence in the new platform. There was no consistent approach to fact and dimension design, making a coherent common data model difficult to build over time.
- Strategic focus: The primary work to date had been lifting and shifting dashboards from Qlik, a necessary step, but one delivering limited tangible value. The opportunity to use the platform for higher-value analytical use cases and AI workloads had not yet been explored.
What We Recommended
Our recommendations were prioritised around what would move the dial fastest and focused on the fixes that would compound over time.
- Establish data modelling principles: Define clear data modelling standards for Databricks and Power BI, moving towards a unified common data model rather than an expanding set of unrelated tables.
- Optimise report performance: Shift computation into Databricks and materialise tables, removing the view-on-view overhead so Power BI performs as it should.
- Enable self-service: Open the platform to business users with proper guardrails, training and data discoverability in place; reducing bottlenecks and building data literacy across the organisation.
- Refine catalog structure: Restructure the catalog with clear domain ownership and introduce dedicated sandbox environments so developers can move faster without risk of conflict.
- Reset business engagement: Move beyond Qlik re-platforming and identify the high-value analytical use cases worth building for, including laying the groundwork for AI workloads.
- Strengthen data culture: Build communities of practice, empower internal champions and invest in data literacy. The people side of adoption is as important as the technical side.
- Streamline ways of working: Establish structured change management to prevent conflicts, and refine CI/CD workflows to eliminate the UAT bottlenecks slowing down every release.
Azure Databricks · Power BI · Unity Catalog · Microsoft Azure
Why It Matters
What they experienced is a common pattern. Organisations make the right platform investments, get the technical foundations in place and then find the expected value doesn’t follow automatically. The gap between a working platform and one that the business uses, trusts, and builds on is almost always a people and process problem, not a technology one.
Closing that gap requires someone who has seen it before across enough organisations to know the difference between a teething problem and a structural issue, and who is willing to say so directly. It also demonstrates something we believe in: the value of an expert perspective doesn’t require months of work to be worth having. Sometimes a focused, honest assessment delivered quickly, without agenda is exactly what moves things forward.
Thanks for reading
If this case study resonates, and you have similar challenges or ambitions – we’re here to help. Book a free strategy session with one of our senior consultants and we’ll talk through where you are, what’s blocking progress, and what a practical path forward could look like.
