Resourcing and Structuring a Power BI Centre of Excellence – Part 4 of PBI CoE Series.

Resourcing and planning for a Centre of Excellence is not as easy as it seems. You might have a business case to fulfil and an agreed scope for the CoE, and in addition, you might have some existing team members and you might need to recruit as well.
This article covers not only the skills, knowledge, and behaviour of your CoE, but also some ideas of how to structure it, so you can deliver the planned objectives.

In my experience, having certain degree of skills, knowledge and personal/professional behaviour has helped my team members and me to be successful in our Power BI CoE roles. There are different traits that are key for such success.

Before jumping directly to lists of skills and characteristics though, we should always remember that (potentially) you are building you Centre of Excellence to fulfil a business case and to achieve the goals that you have defined as part of your CoE Scope.

To read more about Business Case and CoE Scope, please visit earlier posts:

Building the Business Case for a Power BI Centre of Excellence – Part 2 of PBI COE Series. – Smart Power BI

Scoping a Power BI Centre of Excellence – Part 3 of PBI CoE Series. – Smart Power BI

Alongside delivering the agreed scope there is an even greater objective to meet: becoming a trustworthy reference in your organisation for Power BI And Business Intelligence.

In essence, when building your Power BI CoE, you do so to achieve certain goals determined by the case that allowed you to build the CoE in the first place. Those goals areas can be any (or all) of the following:

Architecture and Strategy.

Community of Practice




Instead of breaking down skills, knowledge, and behaviour for each area, we will simplify looking at the most common and critical skills and traits that you might need, regardless the agreed scope of your CoE and the culture of your organisation. However as these both (scope and culture) will affect your resourcing requirements and alongside other factors (budget, availability,…), you will have to decide what is the best recruitment strategy and mix and match, between your needs, existing skills and available ones.

These are the common traits that I found very useful when running a Power BI CoE:

What if nobody from my existing team members or candidates to recruit displays all these traits?

Well, assume that will be very difficult to find all these skills, knowledge, and behaviours in just one person (I am yet to meet the “Power BI CoE Unicorn”), so my approach is to understand the team and their strengths and weaknesses, and make sure that you have a comprehensive representation of all these attributes in the team, not one individual, and if you don’t, then consider self-improvement,  mentoring, and take in account what you have already when facing recruitment.

Now is the time to dive deeper in some of these characteristics.

I would turn this article into a longer paper if I had to develop all the rationale and explanation for all the traits listed before. Instead, I will comment on some important ones, and if there is anything that you would like to know more about, just comment on this post and ask anything you might be interested in.

Let us start with Knowledge. Undoubtedly, technical knowledge is required and the more you have the better, but make no mistake, having outstanding knowledge is not enough to play a good role in Power BI CoE in any situation. I have seen many times how good technical people has provided good technical advice but the wrong one based on the different factors. And technical advice is only one component of the entire CoE, so that alone, does not get you too far.

I must say that combining technical knowledge with Critical Thinking/Lateral Thinking, is a very powerful combination, as opens possibilities to find more efficient, less resistance solutions or even avoiding trouble altogether in the first place, that technical knowledge on its own would not reach just by applying common practice.

Anyway, what that technical knowledge should be? Business Intelligence in general is very important for this ever-changing technology, as you will have to device how it fits strategically. However specific Power BI expertise will be critical to steer and implement that strategy.

Just to be very specific: Kimball DW methodology has been with us for more than 2 decades, and its principles, not only do not change but also are built into different technologies (even outside Microsoft’s toolset). Principles that can and must be applied to Power BI and other Azure components (as it covers ETL and staging).

Solid foundation in Data Warehousing and Business Intelligence concepts is a must have for a PBI CoE member (unless we are discussing a dedicated supporting role, such communications).

On the other hand, specific Power BI technical skills are also very important. However, these skills are candidate for some compromise (for a while), as whilst some technical knowledge is easy-ish to acquire (mainly if there is a good foundation and overall experience), other traits and experience is much harder to come by.

Although this last statement might be shocking, as essentially, what I am saying is that I can compromise a bit on Power BI technical expertise to build a Power BI CoE, let me explain with an example: if I am recruiting for a CoE role, and the candidate doesn’t know about Power BI Premium administration, or has no experience on Composite Models (specifically referring to direct query to tabular models feature), but has solid BI & DW experience, experience with building multidimensional and tabular models and experience troubleshooting variety of issues, not having the mentioned knowledge shouldn’t be a big deal. None of us knew much of direct query to tabular models one year ago anyway.

Specific technical knowledge can be easily acquired by an experienced professional; however, no crash course will give anyone 10 years of experience in developing, deploying and troubleshooting BI solutions.

Understanding your organisation is relevant for a few aspects. You get stuff done by knowing who the best person/team that you can get support from is. The larger your organisation is, the more complex tends to be navigating permissions and processes. You get promotion and reach if you are visible by those that already have influence and authority to take decisions, and with “visible” I mean that they know, trust and understand your and your CoE’s value. Knowing your organisation and its priorities and gaps, can help you identifying opportunities that will lead you to constantly provide value and, at the same time, building on the trust to your CoE.

Whilst knowing your organisation makes everything easier, faster, and more effective, it is something that can be learned on the job. However, if you are building your CoE from the ground up, consider an onboarding plan to get them to know the organisation’s intricacies as fast as possible or find support in some other IT teams within your organisation to help alleviating the gap. Additionally, you can also rely on other IT team/s close to you to provide some additional support during the process.

Alongside with the latter, communication skills, and the ability to plan for effective communications is very important, not only for broadcast type of communications (i.e. when managing your community of practitioners), but also for specific activities with a much smaller audience that are impacted by what you say, how you say it, what you don’t say and level of detail required for the audience. 

Communications are important as wrong ones can cause troubles and unsettlement, and right ones can solve or even prevent future problems. Needless to say, that building effective communication plans, and being generally effective with any communication can be benefited greatly by the earlier point about organisational knowledge.

I cannot turn this article into a lecture about business communications, as it is not the aim and I would not be the right person, but again, based on experience, all kind of communications matter in most of possible scopes for a CoE.

I find report design and user experience key concepts as well, as at the end, the value of Power BI is empowering the business to take decisions based on data insights, and they will do that mainly by consuming Power BI reports. Most of what CoE does (architecture, security, training, guidelines and best practices, governance) is there so information can flow all the way from primary source systems to your users’ eyes (and brains), so if they cannot find and understand the information that they need when they need it, then it’s all wasted.

Based on experience, a significant amount of performance issues in Power BI are caused or contributed by cluttered reports, designed either by the business or with a very generic use case in mind. Having the ability to help turning those extremely busy reports into simpler and more efficient ones, not only to improve performance but also to enhance user’s experience and being able to justify and convince properly that this approach is better, truly helps finding solutions to challenging situations.

Not much has to be said about security; and taking risks on data protection is unthinkable for any business or organisation that want to see the sunrise one more day, and I am not being dramatic here.

When it comes to training, the ability to teach others to use this technology effectively, is pivotal to Power BI platform benefits realisation, as being a true business self-service solution, great value is expected from the business delivering data insights to themselves.

When thinking about training, don’t just limit yourself to programmed classroom or virtual training, but think also that any conversation can turn into training, for example, when during a conversation, you find needed to educate somebody about something. Make these occasions effective and impactful by providing good training (even if it is a 10 minutes one).

Some other skills are more generic and could be applied to many disciplines within a business. They are not only self-explanatory, but very well documented elsewhere.

Despite being generic, they still are critical for a CoE in the same ways as they are for those other disciplines. When running a Power BI Centre of Excellence, you could be influencing or defining strategies which need prioritisation and communication, managing conflicts (technical or otherwise), running projects that have issues and working with or for stakeholders that require your support for their own success. The list of different scenarios where you need all these traits is endless.

Personal and professional behaviour.

However, what sometimes might be underestimated, but very important nonetheless, is your (and your team members’) personal and professional behaviour. I tend to say both “personal” and “professional” behaviour because to me, there is no difference. Obviously, you have professional attributes that you don’t require in your personal life and personal behaviour that cannot or do not want to apply to your professional life, but generally, you are the same person at work and outside. I also like to believe that becoming better at one (sometimes) makes be me better at the other.

Despite I am not a fan of the overused concept of Thought Leader (for some, one annoying form of business jargon), its spirit can successfully drive initiatives such as Power BI Centre of Excellence, mainly when the objective is to become trusted authority in the matter, and a refence in the BI space.

Something that has worked well for me is being honest and justified. We do stuff and take decisions for good reasons, being transparent (as much as the circumstances allow) help others understanding those decisions and actions and genuinely getting others on your side, so to speak. Honesty is a non-negotiable attribute when you want others to offer their trust. Also, in that spirit of honesty, sharing your justifications helps immensely to gain that trust and understanding, leaving aside the opportunity to get feedback and different opinions on those justifications.

I have a strong opinion about risk taking: there are things that you don’t want to be messing around with, but as long as the benefits do outweigh the consequences, and the latter is something that you can live with, there is nothing wrong with taking conscious risks, that is, if the value to the organisation that could come out of taking it is worth. Do not use this as a blanket to do silly stuff, and always try to be objective assessing the benefit and consequences.

The kind of risks that I am talking about are more related to loss of credibility than anything else. I have a theory: you may try and fail or try and succeed. Failures might affect negatively to your credibility, successes will not always contribute to higher recognition, but a constant approach of not being afraid of failure, for the sake of your organisation, being honest and clear about benefits and risks, seeking always the benefit for others and your organisation over your own reputation, provides a different level of credibility, that I particularly like.    

Last, but not least, any aspect that helps improvement should be built in anyone’s personal and professional life. Accepting criticism (even when it is not constructive – yeah, I said that) and trying to assess the truth in it and having the ability to put under question your own behaviour, skills and knowledge are some of the most powerful tools immediately available to you to always improve in area that might serve you.

Organisational Structure.

I am frequently asked about how I would structure the Power BI Centre of Excellence, and despite that I don’t generally like being too prescriptive, I see the value of having a base schema as a starting point.

Obviously, the CoE still needs to be structured based on the scope agreed and on what areas it will deliver value to the organisation. A basic structure would look like this:

The rationale for this approach is as follows:

  • Each main area deserves dedication, including a manager (who is accountable, alongside the Head of CoE), which should help with:
    • Right focus and making sure that the things that matter get actually done, via accountability.
    • Caring about the relevant skills and knowledge required for such area.
  • Community lead might not be deemed as a time-consuming activity and can be supported by other CoE members. This role can even a second hat for other/s team members.
  • As CoE might not have all the resources needed for each area, or not all areas require full time employees, it is good to have a pool of resources that can change hats.
    • This would also help with team members’ motivation.
  • Head of Power BI is also ultimate responsible for all decisions and actions taken by the CoE. Head of CoE should:
    • Define and lead the strategy.
    • Own the plan and execution of the CoE and make sure that delivers value as promised.
    • Know when to change direction of travel and re-prioritise objectives.
    • Support and guide the team.

Anyone implementing a PBI CoE might have their own organisational requirements and/or preferences, therefore it is absolutely possible to make changes to this structure, however it is important to consider some principles common to any team:

  • Ensure clear accountability and responsibility.
  • Ensure that the right people are in the right role.
  • Provide good and timely managerial support and guidance to any team member.
  • Care for relevant skills and knowledge.
  • Ideally, provide growth opportunities as well.

Unlike other business functions, a Centre of Excellence requires a set of attributes to achieve ambitious view. The ability to adapt to changing business needs, to absorb technology evolution and to learn and improve are pivotal to success, requiring a combination of traits that needs to be caressed.

Power BI Centre of Excellence Series Articles:


  1. Hi Alex

    Thanks for the very detailed blog series on CoE, Unilever is said to have the WORLD’s biggest Power BI deployment,

    Could you share what does your Power BI Premium (SKU) look like, I mean how many P3 SKU’s you have


    1. Hi Zahid,

      I believe that Microsoft’s implementation is the largest one. My company’s is one of the top, though.
      Instead of answering your question with a number, I am going to ask you another: what is the intended use for your Premium state?
      Let me explain.

      If I answer your question with 5 (P3s) or 50, it wouldn’t mean much. What matters is your organisation’s approach to licensing (do all employees have a Power BI Pro or O365 E5 subscription?) and your approach to dataset hosting (is your org leaning towards PBI Premium instead of Azure Analysis Services?).

      There are also other relevant questions, like: do you have a large external (to your organisation) user base without PBI Pro?

      With these questions in mind, and given your current usage, you might think that your Premium state might grow.

      Because, your starting point (how many Premium capacities you have, and their sizes) is important, it is not critical. You can start small and purchase as you see that need more capacity.

      Also what is very relevant is how to manage such capacities and avoid waste, in the long run, caused by inefficient solutions that consume more resources that what they should – this is a whole world on its own…


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s