Data Governance Operating Models

- Flexible Operating Model

Data governance is an essential foundation for organizations looking to create business value from data. It creates a structure that supports collaboration and analysis of trusted data. However, establishing effective data governance can be very challenging. Data governance practices require enabling technology, and that technology must include a flexible operating model that allows organizations to design governance programs in a way that suits their unique needs. 

An operating model outlines how an organization defines roles, responsibilities, business terms, asset types, relationships, domain types, and more. In turn, this affects how workflows and processes function; it affects how organizations function around their data. 

An operating model is the foundation of any data governance program. An operating model helps establish a corporate governance structure by defining corporate roles and responsibilities across different lines of business. 

- Types of Operating Model

There is no one-size-fits-all operating model in data governance; an organization must adopt the operating model that suits its needs. The operating model must balance two extreme principles, centralization and federalization. Organizations can adopt one of these two types, or even embrace both principles. Organizations can adopt one of two operating models, centralized or federated, according to their needs.

• Centralized operating model: A central authority decides the rules for how data in an organization is managed. This central body defines standard processes for implementing data governance principles, such as how key data elements are defined and how business terms are approved. When individuals and teams perform data governance tasks, they must adhere to centrally defined processes.
• Federated Operations Model: This is when there are multiple permission groups. Organizations may adopt this model if different teams have different data governance needs.

Both types have advantages and disadvantages: 

• Centralized operating models are simpler and easier to maintain, but they can leave some teams feeling constrained by a central data governance body. 
• A federated operating model allows different teams to adopt data governance at different speeds, but at times it can become more complex or difficult to maintain. 

As a result, organizations often adopt a hybrid model tailored to their needs, balancing the benefits of centralization and federation.

- Practitioner-led Bottom-up Data Mesh Approach

As the number of data users and consumers continues to increase, having a few people (data stewards or engineers) accountable for data governance is not a sustainable approach.

A decentralized, bottom-up data governance framework that holds every data creator accountable for data governance is the way forward.

An example of a decentralized, community-led approach is a data mesh. Data mesh design proposes a federated computing governance model where each organization is a federation of business domains. Domain owners fully manage the data they create.

However, each domain still follows a set of global (or federal) rules on data definitions, standards, processes, and discovery

- Data Mesh

Data mesh is a strategic approach to modern data management and a way to enhance an organization's digital transformation journey as it centers on delivering valuable and secure data products. 

The main goal of Data Mesh is to move beyond traditional centralized data management approaches utilizing data warehouses and data lakes. Data Mesh emphasizes the idea of organizational agility by empowering data producers and data consumers to access and manage data without delegating tasks to data lake or data warehouse teams. 

Data Mesh's decentralized approach assigns data ownership to domain-specific groups that service, own, and manage data as a product.

[More to come ...]