So what is the point in all this Power BI Centre of Excellence stuff about? We write article upon article about the benefits of a Centre of Excellence, but I guess we are guilty of assuming that you already know why a Centre of Excellence is so valuable and a core part of any successful Data Strategy (to be honest even if you don’t want to use Power BI – but we cannot imagine why that would be).
In a traditional reporting landscape, there may or may not be a central reporting platform, but any updates required within it must go through a lengthy approval process and often require budget as there is a cost for even a “simple” change. This is buffered within the organisation by an extensive network of domain-specific people who produce specific reports within a subsection of your organisation. These reports are often as important, if not more so, than the enterprise solution, although few people will acknowledge that truth. This way of working leads to an almost secretive reporting environment as people do not want to get in trouble, and they don’t like to admit where the content has come from. These are known as Shadow Report Functions. Those involved in Shadow Report Functions are often almost Evangelical, seeing any comments about the function as a criticism and being very closed-minded towards changes. Frequently, they have been “burned” when centralised changes have been made that did not involve them, leading to them getting in trouble while the centralised function scrambled to restore content, they were not aware existed.
Does this sound familiar? These Shadow Report Functions are also often considered “Silos”, where knowledge and information are hoarded; after all, the saying is “Knowledge is Power.”
Let us start the next part by discussing what the elephant in the room is often in terms of business discussions. Data is data. However, different roles exist within your data landscape. The simplest one is data engineering, which is the function that takes data from source systems and then prepares it for analysis in a standardised way. Looking back at our Back2Basics video series, this would be the ETL steps; broadly speaking, that is “all” it ever is. The changes are to the complexity of the ETL process and where data is “landed”. In our video, we land all our data in a Power BI Semantic Model. Confusion comes from the other side of Data, this side is perceived as “forward facing” and is made up of Data Analysis and Data Science. These are very much misunderstood and understanding what they are makes understanding why you need a Centre of Excellence obvious.
Data Analysis is part of Business Intelligence, at the most basic level this is about utilising known business and data rules to produce reports to track the specific KPIs that drive the organisation.
Data Science is a scientific function. The role of science is to explore and find new knowledge (insights) about the business to confirm/invalidate/identify business rules and logic. Data Science should not be directly end-user-facing as it can use not fully verified business logic. Data can be presented and shared but must always be heavily caveated. Data Science is very much the area of Data that gets the attention at present due to its involvement in the development of Machine Learning models.
Business Intelligence (and Data Analysis) seems almost to be a dirty word now. Still, if you ask Ross, he will describe himself as a Business Intelligence professional because you do not have analysis without Business Intelligence (or Data Analysis). By providing that backbone, you can produce new content and compare it before making it mainstream by bringing it under the auspices of Business Intelligence. After all, a machine learning algorithm running in isolation is useless; you need to know it has been validated and approved.
So, when we understand that your Data Analysts provide value by using your standardised, approved, and validated data, we can see the value that they bring. Your Data Science function should be looking to undermine the existing reporting by checking whether assumptions are correct.
In his book “Think Again”, Adam Grant talks about our modes, our Shadow functions work in Preacher or Prosecutors, looking to convert people to the correctness of our Shadow functions or prove people wrong for saying it is not perfect. These are the attitudes of entrenchment. Data evolves when we are willing to let go and accept that we may not know everything about our business. We have to adopt the mode of a Scientist. Freeing our minds of the shackles of certainty and embracing that core sentence that underpins every “I don’t know what I don’t know”.
Success in Data only comes when we can align People, Process, and Tools; the Microsoft Power Platform is a tool that will enable you if your process and people also agree to it. If not, the Power Platform will still deliver, but it will deliver the same legacy-style solution you wanted to get away from.
Only when you understand that “You don’t know what you don’t know” can your people come around to the necessary process changes. Those process changes are what will be delivered to your brand-new Power Platform instance transformation. Suddenly it is clear what is needed.
The Power Platform alone will not save your organisation. Teams that have become entrenched cannot deliver organisational insights purely because they are traditionally aggressively kept “in lane.” Requiring lengthy approvals and budgets for “simple” work only delays insight generation and makes for a slower, less cost-effective organisation.
When you understand that success does not just come from buying the tool, implementing a Centre of Excellence provides the flexible framework that an organisation needs to realise the full benefits. Every company has someone who will produce insights using a spreadsheet. The Power Platform can empower them and allow their skills to shine for the organisation. For some, that may be a change in career direction; for others, they remain where they are in the organisation but now work “in the light,” as it were, advancing the business and being able to support other teams through their involvement in the Centre of Excellence. Finally, because we stop trying to prevent the business from self-service, the ability to explore the known landscape of a data model can be made available to many more with relatively low-cost licences. Suddenly, simple requests can be completed either by self-service or with simple support from Centre of Excellence members. Very quickly, a periodic cycle that allows significant changes to the model to be made becomes clear. That process is recursive as more dimensions and data are added to the model, so further low-cost yet high-value insights start to come out of the model.
This structured environment also empowers Data Science. With a stable core to evaluate hypotheses, assertions can be tested, models trained, and preparations for the next stages of analysis made.
It is true to say your Power Platform deployment has its lowest business value on day one.
Geordie Consulting can provide the expertise and guidance necessary to navigate the complex landscape of modern Business Intelligence and Data Analysis. By partnering with us, you will gain access to a team of professionals adept at leveraging tools like the Microsoft Power Platform to drive your business forward. We understand that successfully implementing these tools requires more than just technical knowledge; it demands a holistic approach that aligns People, Processes, and Tools. Our consultants will work alongside your team to foster a culture of continual learning and adaptation, ensuring your data analysts and scientists can thrive. We will help establish a Centre of Excellence, empowering your employees to harness the full potential of your data through self-service and collaborative efforts. This will enable your organisation to rapidly respond to changing business needs, uncovering valuable insights that propel you ahead of the competition. With Geordie Consulting, your journey towards a data-driven future is not just a possibility but a certainty.
