What is the appropriate data and report ownership for Power BI?

   Check list of all posts

Context

The initial motive of answering this question is from my consulting job as a DW/BI consultant, where I've spent most of my time developing Power BI solutions. While the IT department owns everything for Cognos Analytics, it may not be the same for the Power BI.  Power BI is the leader of Self-services BI. As a result, there is no generic answer for Power BI, and it will totally depend on the company context.

I will share three cases using a different ownership structure. In addition, the data flow architecture is also different to accommodate data and report ownership.

Data and report ownership

I have borrowed the diagram below from Microsoft, indicating clear definitions of each ownership mode.

If you pivot the table above, you will get three modes below.

Case 1, Enterprise BI:  IT department owns both data and reports

Case 2, Business-led self-service BI:  Business unit owns both data and reports

Case 3, Managed self-service BI: While IT department owns data, the business unit owns reports

Case 1, Enterprise BI

Enterprise BI is a centralized approach in which all content is owned and managed by a centralized team. This team is usually IT, enterprise BI, or the COE.

This client is a technology company; it provides software to more than 5,000 lead companies worldwide. Power BI is white labeled as reports / Dashboards for customers. To ensure the quality of reports, the technology company has applied DevOps with CI and CD approaches. The diagram below illustrates the overall data flow architecture. 

There is minimal logic in the Power query, which is removed from dataflow for clarification. Please check blogs such as for reference (How to achieve the Power BI shared datasets architecture (Thin Report, Golden Dataset; How to build a data warehouse without ETL and database using Synapse serverless for structured data). Image if we use this process to develop Power BI solutions for internal business units, it will cost massive effort, but can't deliver the result in the short term. 

Case 2: Business-led self-service BI

With business-led self-service BI, creators and subject matter experts own and manage all content. Because responsibility is retained within a business unit, this strategy is often described as the bottom-up, or decentralized approach.

This client is a business unit in a bank that provides liquidity stress test analytics. This business unit drives the projection data,  and develops reports directly on the Power BI portal. There is no data warehouse, and Power BI developers create all power BI datasets by working with business SMEs. The Power BI dataflow was used to develop conformed dimensions, such as bank calendars and shared dimensions for all different data streams. This case is pure business-driven and achieves results very quickly. The diagram below illustrates the overall data flow architecture.

We have used Power BI dataflow and dramatically used Power Query for all logics for Power BI datasets. Please check blogs such as for reference (How to calculate running totals by group in Power Query;How to create Power BI tabbed visuals with or without common slicers).

Case 3: Managed self-service BI

Managed self-service BI is a blended approach. The data is owned and managed by a centralized team (such as IT, enterprise BI, or the COE), while responsibility for reports and dashboards belongs to creators and subject matter experts within the business units.

This client has SQL data warehouse, which is well maintained by IT department. The objective is to allow the business unit to achieve more agile, responsive and business-driven analytics. We can't apply for Business-led self-service BI, as the business unit will face data issues or can't get a single version of the truth. in addition, the business unit should not directly retrieve data from raw source either. We also can't apply for Enterprise BI, as the business unit can't wait for IT department to handle their requests with a traditional SDLC. The idea is to rebalance the workload between IT department and the Business unit. Technically, the job of IT department is to provide "golden data," while the business unit side is to create the best Power BI visualizations. The diagram below illustrates the overall data flow architecture. We do need to build some logic in Power Query while we Power BI dataflow as we have a data warehouse.