ABSTRACT: An operating model for data & analytics is critical for aligning resources across the enterprise and balancing the needs for agility and governance.

This is the first of a multi-part series on data & analytics operating models. 

Most data leaders focus on architecture or data governance when trying to increase the value of data & analytics in their organizations. In my 30 years of experience, however, the most important thing for data leaders to focus on is their team’s operating model. 

According to the Operational Excellence Society, an operating model is “…a visualization (i.e. model or collection of models, maps, tables, and charts) that explains how the organization operates so as to deliver value to its customers or beneficiaries.” An operating model shows who does what for whom and how. 

In the data & analytics field, an operating model is critical for aligning resources across the enterprise. Specifically, it closes the gap between the enterprise team and data users, eliminating stubborn data bottlenecks; it empowers users with just-in-time training, support, and coaching to foster greater levels of self-service; and it propagates standards, templates, and best practices to minimize the proliferation of spreadmarts and data silos. 

High-Level Framework

At a high level, a data & analytics operating model is simple. It divides the organization into technical and business domains with a hybrid unit connecting them. Technical knowledge is highest among technical teams and lowest among business teams, while domain knowledge is highest among business teams and lowest among technical teams. (See figure 1.) 

Figure 1. High-Level Operating Model for Data & Analytics

Technology teams

IT Stakeholders (BLUE). The enterprise data & analytics team partners with various IT departments to ensure uninterrupted delivery of data & analytics solutions and services. The chief data officer (or head of data & analytics) needs to maintain tight partnerships with infrastructure, enterprise architecture, project management, and security teams. 

Enterprise Data & Analytics Team (BLUE). The enterprise data & analytics team (in blue) manages the enterprise data environment. This primarily involves building and maintaining an enterprise data platform and enabling business domains to build their own data products and solutions. Other than dogmatic data mesh adherents, most enterprise teams also manage shared data (i.e., a data warehouse and data lake), build complex enterprise and cross-functional solutions, and manage metadata and the data catalog. 

Hybrid teams

Domain-Based Development (PURPLE). This entity bridges the technical and business teams and is the key to an effective operating model. The purpose of this group is to dedicate technical experts to a single domain so they can build local solutions quickly and effectively. At the same time, these “purple people” adhere to standards, processes, and best practices established by the enterprise to ensure data alignment across domains. These teams balance agility and governance, giving business users what they need without creating data silos. There are many ways to design this function. Some organizations use tiger teams or centers of excellence, while others use business analysts or “spanners”. Data mesh adherents advocate for data domain teams. 

Business teams

Business Domains and Stakeholders (RED). Business domains align to high areas of data demand from business stakeholders. A business domain could serve a single stakeholder (i.e., marketing department) or multiple functional areas that converge on a business opportunity, such as customer journey, supply chain, or digital channels. Data-hungry departments have one or more data analysts who answer ad hoc questions from business users using ad hoc query, data preparation, visualization, and perhaps autoML tools. Data-hungry domains support a development team (i.e., the purple team above) that builds and maintains local data products and solutions that are more complex than what a data analyst might deliver. 

Role of the Chief Data Officer 

Even in a small company, an operating model can encompass dozens of individuals and roles spread across multiple departments and regions. This puts a premium on the organization’s data leader to establish the connections between blue, purple, and red entities to foster alignment and balance agility and governance. The leader does this by managing how standards, processes, and best practices trickle down and domain knowledge trickles up. 

Next In the Series

In our next blog, we’ll explore linkages between entities in the operating model. We’ll describe reporting models and types of information that trickle down and up. In subsequent blogs, we’ll define the teams and roles in each tier of the model and how they interact to deliver timely data and insights to meet business needs.