Skip to content

Building an Analytics Team for Your Organization Part 3

In a previous blog post, we discussed the foundation of building an analytics team including the four key elements: Leadership & Governance Structure, Scope of Services Charter, Teams and Roles. This blog installment will center on the design of centralized and decentralized models and will also touch on the interdependencies and how firms manage work across organizational lines.


For the purposes of setting forth at least the basic principles for structuring an analytics function, consider the following organizational models.

  • The centralized analytics operating model includes general analytics practitioners as well as practitioners specific to corporate functions, such as human capital analysts to support the Human Resources function and shared services analysts aligned to the operations and shared services functions.

  • The decentralized analytics operating model includes general analytics practitioners within a central group aligned under Finance, IT or a stand-alone Analytics department. Any process or functional-specific analytics practitioners resides within their respective home organizations – Finance, Legal, IT, Operations & Shared Services, Marketing, Research, Customer Services, etc. In fact, there is no true organizational model for decentralized, but instead, the required roles to deliver a complete, in-house analytics capability are parsed throughout the various areas of the business needing an individual with data analysis or reporting skills.


The key difference between the Analytics Delivery Organization and the Analytics Functional Team is the top-down reporting structure. The Analytics Delivery Organization reports directly to the CAO. The Analytics Functional Team is deep within one of the corporate functions, such as Finance, Marketing or IT. Therefore, a separate org chart is not represented here for the Analytics Functional Team.


The key ingredient to this model is the location of the business analysts and functional experts. As noted, they are not reflected within the central organization. Instead, each corporate function must assess the need for the level, type and scope of analytics to be produced within its business processes and then staff accordingly.

A recommendation to consider concerning this organizational planning and staffing concern is to outline analytic capabilities within the overall functional plans and / or road mapping work for each quarterly and annual cycle as applicable. Reflecting the analytics capabilities within those plans rather than treated as a separate exercise or plan may led to gaps or inconsistencies in planning assumptions as well as neglect to consider and address how those analytics capabilities must be integrated within the day-to-day operations of the group.


One could focus an entire article or research study on how an Analytics organization can effectively integrate with its business partners, supporting corporate functions and any other external stakeholders such as vendors. However, for the purposes of this article, I will cover the highlights here as a matter of explaining that it is essential to have clearly defined working relationships in any analytics operating model. So here are some of the primary considerations for managing work across organizational lines:

  1. In all of the analytics delivery work whether it is a technical platform being implemented or a new set of analytic models, set forth a clear project plan with dependencies defined for all work streams and all deliverables captured with one clear owner.

  2. Ensure that org charts clearly show dotted line relationships, such as any teams that exist within IT but provide day-to-day support of the analytics technical infrastructure. The dotted line relationships should be clear in terms of which roles interface with each other and at what levels. Are the relationships between the two groups driven at the manager or director level? Who is responsible for overseeing that work is produced on-scope, on-time and on-budget when both teams are involved?

  3. Establish clear budgetary guidelines as well. Are overhead costs distributed across more than one department, such as costs for licensing or vendor services, or are they assigned solely to the central Analytics function?

  4. Define a regular cadence of status meetings involving all of the involved organizations to ensure that the right level of oversight, day-to-day guidance, and staffing is in place and that the projects are being delivered on-time according to any roadmaps or project charters. An executive steering committee with the right level of commitment and involvement can make a tremendous difference to the quality and overall success of an analytics function as well as for any internal function or project-based organization for that matter.

  5. Ensure that project reviews, stage gate checks for deploying new solutions, and any quality assurance review or audiences account for joint sign-offs across the multiple functions or teams involved; giving the senior executives accountable for the success and day-to-day results of each team a voice to either identify a success and approval to move forward or conversely, a voice to halt a project or retrench if needed.

One way to take the concepts outlined here concerning a well-designed analytics operating model is to seek input from analytics practitioners across various organizations, industries, and organizational types [centralized operating model versus decentralized operating model].

If you are interested in providing your perspective on this topic, I would ask you to contact us and note whether you are interested in assisting in a research effort to further explore this topic and to develop some guidelines and best practices for analytics organizational design principles. I plan to continue to further explore this topic and to incorporate findings and recommendations from practitioners across the industry as we need to continue to seek more definitive answers and guidelines for organizations setting out to expand or establish a dedicated, in-house analytics function.