Analytics Centre of Excellence – What’s the value?

Benjamin Wright-Jones, Solutions Architect
Contributors: Simon Lidberg, Michal Hlobil, Ken Collins, Manoj Kumar
Reviewers: Carla Staeben

The rising interest in data and data science has led to a natural interest in the creation of business functions to manage and accelerate the value of information and analytics. From a Microsoft perspective, we have observed an increase in these discussions given the rise of self-service business intelligence through tools such as Power BI. Organizations want to empower end-users with tools such as Power BI whilst also ensuring integrity and consistency at the core.

The term Centre of Excellence (CoE) or Competency Centre (CC) have been used since early 2000 by Gartner (Morello, 2002) and others to describe a cross organizational group responsible for a specific function such information management, with an ultimate goal to reduce time to value.

The CoE function has been used with great success by many companies and organizations when implementing ERP and BI systems (Howson, 2009) (Dresner, et al., 2002). Specific to analytics, you may hear such a function referred to as a Business Intelligence, Information Management or Analytics Centre of Excellence; it can take many guises within organizations.

The main purpose of a CoE is to establish, identify, develop, and harness cross-functional processes, knowledge, and expertise that have tangible benefits for the business. They serve as channels through which project managers, practice managers, business customers, and line managers can fulfil projects, teams, and business initiatives. Most important, COEs continually generate and refresh knowledge, competencies, practices, and skills with the goal of informing and guiding others working in the same business domain.

The following diagram is based on our discussions with enterprise customers and depicts some of the COE capabilities.

Given the current focus on Big Data and the Internet of Things, governance is a crucial part of any CoE; indeed, it is so important that it will be the subject of a separate blog posting. However, it is worth saying here that governance doesn't have to mean large, complex and expensive programs of work. Rather it can and should be tackled in in small focused efforts where there are specific challenges rather than a large corporate project.

I was reluctant to list governance as a capability given it is a 'loaded' and often 'scary' term, suggesting expensive programs of work. I would also be inclined to avoid using the term 'governance' but need to reflect this capability in the CoE model.

Some of the main drivers we have observed (for establishing a COE) include:

  1. Recurring costs of current solution upgrades
  2. Increase adoption of analytics
  3. Standardization of analytics tools across the organization
  4. Being able to accelerate the go to market of new analytics projects
  5. Functional limitation of the current solution
  6. Rationalizing analytics investments with Microsoft Offering
  7. Improved data management, data quality and reporting
  8. Better collaboration between business and IT
  9. To make analytics accessible to the business

It is also worth noting that such business functions are not technology specific and have to span a multitude of platforms and cloud services e.g. Azure Data Warehouse or Azure Data Lake. Interestingly organizations are looking to software vendors to help guide them with information management programs (primarily due to their experience in the data domain).

Roles and responsibilities are also very relevant and as such we typically recommend executive sponsorship as mandatory, then a CoE lead/director along with one or more of the following: program manager, project manager, business and solution architect, developer, data scientist.

It is often a good idea to include a data scientist role in the CoE. Many companies do not have the need to have data scientists in the organization but instead the CoE data scientists can work with domain experts from the different parts of the company to help them apply advanced analytical methods on their data. Another reason for having the data science function in a central organization is that it would ensure that the data scientists can collaborate successfully and exchange knowledge without being constrained by organizational boundaries.

We also recommend a role to monitor performance management which is crucial in reporting the return on investment of a CoE function to the executive sponsor and/or senior management team, for example, where the CoE has positively impacted projects through faster delivery, re-usability of lessons learned or knowledge generated from other programs etc.

Microsoft has a vast amount of experience helping organizations gain insight and value from their data. Some of these programs of work often start with a discussion about specific challenges which can evolve into understanding business maturity in analytics and the organizations capability which is where the conversation naturally leads to the benefits of a CoE.

Business Intelligence and Advanced Analytics with Microsoft Services

The Microsoft Enterprise Strategy Program

This posting is provided "AS IS" with no warranties, and confers no rights. Use of included script samples are subject to the terms specified on