The Functional Centre of Excellence

Over the last few weeks, we have shown you that there is a significant benefit for a company that can adapt and embrace the principles of a Centre of Excellence in its Data landscape. We have discussed what that means for you as a business user and how a Centre of Excellence brings much greater latitude in interacting with data within your organisation. The functions served by a Centre of Excellence are multiple and will not be carried out by the whole team; instead, they will work together to ensure delivery in the long run. Considering the tasks of your functioning Centre of Excellence from Source Data to visualisation, all data sources must first be documented, including refresh frequency, start time and expected refresh duration. Schemas are a challenging concept at this point. While your Semantic models should be optimised around what is required, we know that will change - it may be that there is a lot more information available in the data source than is brought through. Remember, Data Source ?? Power Platform / Cloud Database is not the same as what is in the Semantic Model. It is advised to capture the points of interaction, at least those that could be primary key or foreign key columns, which are entity table names at a minimum.??

The process of building semantic models is ongoing, and once it starts, it should never stop; models will always require updates, and the frequency of updates will change as the model matures and becomes more valuable to the business. This cycle is the most crucial element of a well-functioning Data Centre of Excellence. This is, after all, managing how most people will interact and access corporate metrics. The decision of what Data Scientists within your organisation can access is one that each organisation will need to discuss amongst themselves; critically, Data Scientists should be part of the Centre of Excellence even though, by definition, their role is not about producing metrics for the organisation. However, the explorations they complete should add value to Semantic Models as they can be packaged into additional metrics and then wrapped into new Report packs. Semantic Model(s) serve all the line of business reporting requirements of your business; this means that they must be able to continue to grow and evolve with the requirements of your business. Understanding that, we can see how we must ensure a clear start point mapped for the data flowing into the model. Then, we can use tools like Vertipaq Analyzer to complete your technical documentation.??

A Power Platform Project would typically produce a small number of reports and dashboards to create one or two Apps for use by the key stakeholders; with that done, the Project Team will then rapidly start disconnecting. This is, however, where the work of visualisations starts; this is how value is brought into a business. Anyone with an appropriate paid subscription can create new content based on the Semantic Model using Power BI Desktop (available from the Microsoft Store) or Excel (full install only). This means that when at lunchtime on the first day (yes, it is often that quick) a “new” requirement arises, the Project mentality of “Well, that wasn’t what was requested” simply is not appropriate. Instead, Power BI gives two options: either all the information is there and a new report can be created using the different spin on the data, or the core values are not within the Semantic Model, meaning that the requestor has two options, firstly they can reframe the request in the context of the insights currently available or they can wait until the update can be made to the Semantic Model, either by adding the additional values from the available source or going back to the source owners to have the source updated. Pause now and consider that new requirement.??

The agility of the Power Platform and the dynamic nature of DAX means that many new requests are not new; instead, they are just a different way of presenting the existing semantic model; even new Measures can be added to the Thin Reports that you will be creating, this allows new metrics to effectively be “Staged” pending the next periodic Semantic Model update (should they need to be made available to all). Secondary Reports can be made available in under an hour (in terms of working time, there may be a delay due to staff workloads) even if the requestor must reframe the request in terms of what is available now, for many that is far preferable to having to wait days or longer to get anything. Again, the platform agility is showing additional value. The final option is the one that makes the initial launch challenging. It is not uncommon for the first week for near-daily updates to be pushed through to meet all the requirements that were “missed” during the initial build. That period can lead to you losing faith in the Centre of Excellence concept; however, this is the early stage of the business's acceptance of the model. The addition of new metrics and dimensions serves to increase the value that the platform provides to the business. A critical function of the Centre of Excellence is to take that pragmatic approach to update the Semantic Model, as the frequency of model refresh is directly related to the value it provides, meaning that the more infrequent the updates, the more valuable the model should be (remember to check usage figures on this – it could just be no one is using the model). After the first rush for “urgent” updates, a refresh cycle and “front door” process should be established (something that the Power Platform does very well). Starting with a weekly refresh cycle and then extending that to fortnightly, monthly, quarterly and finally annually. Reviewing requests and working through the enhancement triage process to provide consistent delivery while allowing new work to be completed (remember this may also be updating the Semantic Model).?

You must understand and recognise why we recommend considering the actual reports and dashboards within the Power BI Service as disposable. When you see a production environment with 50+ reports touching the Semantic Model, it becomes vital that we ensure that the model is the point of stability. Measures should not be deleted or radically changed without consultation. Equally, measures can be updated so, for example, if a change is required to how your “Net Profit” is calculated, that can easily be updated once centrally and have all the child (Thin) reports automatically updated. The last type of improvement that the Centre of Excellence will do for your business is to work improvements into the semantic model as part of each update cycle. The Service performed by the Centre of Excellence is therefore one of Continual Service Improvement.??

We have explored the concepts that must be understood to achieve success in terms of what a Centre of Excellence must achieve. Activities will come in for updates and this is as it should be, in fact if the “business” goes “quiet”, that is the time to worry! A steady stream of enhancements interspersed with new functionality/data along with performance tweaks goes a long way to ensure continued success of your Data Programme. Members of a Centre of Excellence perform a vital function for your organisation and their success will always rely in particular on the continued support of the Senior Leadership Team of the Business. The biggest challenge faced by a Centre of Excellence is always the first releases. This is a time when the still relatively new Centre of Excellence function can implode as it is unable to address the “urgent” nature of that first launch and is why you may find it beneficial to seek external assistance to support your team and their operation while they are still finding their feet.??

Geordie Consulting offers a Centre of Excellence as a Service to help you embed the Microsoft Power Platform to provide you with agile, cost-effective data solutions. Our experience goes far beyond the implementation of Data Tools; we also specialise in integrating data from across business applications into a single, manageable platform. Our services include enabling business users to develop custom reports using Power BI and Excel, leveraging Semantic Models for consistent data analysis, and enhancing user self-sufficiency through Copilot. We aim to ensure your data agility to support a data-driven business. We provide comprehensive documentation and support to maximise the business value provided by your data. Let our insight unlock your insights.?