Maturing Analytics Through a COE
Data and Analytics are changing the landscape of business, allowing for unprecedented insight and the ability to leverage today’s data to predict possible future outcomes. This new status quo is rapidly changing the very way that businesses operate, with each organization trying to set up the needed components to push past their competitors. However, this rapid push to enable analytics comes with its own sets of problems, with business units making use of differing tools, having differing levels of technical skills, information silos that bar cross-functional teamwork, and differing sources of truth for the same metric.
At Slalom, we commonly come across these problems at our clients and have spent years searching for a method to address these and many more issues. To us, one of the simplest and most robust answers to these problems is an Analytics Center of Excellence (COE). These problems might have technical ramifications but at their core, they are organizational issues that can be addressed through the proper organizational structure and guidance. While it is possible to scale an enterprise’s analytical capability without a COE, we find that a COE helps address common issues and speeds the organization up the analytics maturity curve.
This curve represents how an organization can mature its analytics capability, with the bottom of the curve representing budding analytical practices, to the top showing a truly data-driven organization. For the most part, organizations fall somewhere in between and are now trying to scale to the next level, focusing on increasing the ROI of their analytical pursuits. Growing in maturity is one of the more challenging aspects of growing an organization’s analytical practices, as there are few levels to plateau on before having embedded analytics. A COE can help smooth that transition and speed the time to scale up the maturity curve.
Much of the difficulty in scaling up this curve and how a COE helps comes back to core technical tenets and organizational best practices. Analytical projects require a platform and architecture that supports them and that should be utilized across all business units. This helps to smooth issues with data security, model compliance, and the maintenance of various analytical projects. Having all analytical governance residing in the COE results in less redundant work scattered across an organization, allowing developers more time to focus on their day job and less on work to ensure compliance.
The increase in efficiency can be attributed to how a COE will evolve as it moves up the analytics maturity curve. Most COE’s begin as an effort to combine analytical assets from across the organization under one leader. At this point, most analytical projects that have been implemented rely on multiple systems and lack a cohesive data source, in a worst-case scenario the output of these early projects could conflict with one another. It is at this early stage, that we would expect to see a COE Lead work to define a source of truth for the organization and begin to organize a backlog of projects for the business units they look to support.
As the organization moves up the curve into “Competency” we would then begin to expect to see the team growing, possibly adding the light blue titles in the above figure. It is at this point that we expect to see an increased ROI from the COE as solutions are being stood up, with the COE Lead acting as a liaison with the appropriate security and compliance folks. Solutions are being built more robustly and we could see an increased focus on automating solutions, resulting in developers having more capacity for active development.
Finally, the organization reaches full maturity. From a COE Organization structure, this might look like having dedicated assets for Architecture, Data Engineering, Data Science, Dev Ops, and Visual Analytics. The organization leverages analytical solutions to influence operational decision-making at the highest levels. At this point the focus of the COE is working on fully automating solutions, incorporating robust Dev Ops practice, and ensuring solutions are automated. It is at this stage that the organization becomes truly “data-driven” and able to incorporate best-in-class analytical decision-making into their operational processes.
For more information on setting up a COE, please follow the link to a webinar recording on this. For help diving further into this topic, please contact hellodallas@slalom.com for details on how we can help grow your analytical capabilities.
