Use Cases Need to Drive Priorities When Planning and Investing in Data Fabric and Data Mesh Technologies

To capture the benefits of more adaptable, flexible, and automated infrastructure to support evolving BI needs better, organizations must modernize several technology capabilities. Most important is the need to deliver a “layer” that helps translate the myriad of physical and distributed data sources in most organizations into a more business-friendly means of accessing and manipulating that data. Visionary data leaders increasingly invest in this direction, implementing capabilities of an Active Data Architecture™ (see the Research Insight “Data Mesh and Fabric Confusing? Data Leaders Need to Unpack Next-Gen Data Architectures”).

An Active Data Architecture supports a platform-independent layer that sits between physical data stores and points of data consumption. It is comprised of various data management capabilities including virtualized and distributed data access, data governance, and security. At its core, Active Data Architecture is an abstraction layer translating business and physical structures. Its architecture dynamically optimizes for performance, scalability, and cost management.

The concepts of data mesh and data fabric are sometimes associated with Active Data Architecture. They involve managing distributed data, enabling consolidated data views, and provisioning data for consumption by processes and applications. Data mesh links distributed data sources and enables these capabilities in a pre-programmed, practitioner-managed, and manually optimized fashion. Data fabric builds on these capabilities and adds elements of automation to help make the Active Data Architecture dynamic, self-organizing, and continually optimized.

Building an Active Data Architecture is a cross-functional investment that requires engaging multiple teams, tapping into existing budgets, and expanding centralized investment in key areas. It requires data leaders to carefully plan, prioritize, and sequence modernization efforts to maximize value and ensure success. While advancements in technology capabilities sit at the heart of an Active Data Architecture, data leaders also must identify use cases that can directly impact strategic business outcomes.

Successful data leaders seek high-priority data-product use cases, which require a modernized Active Data Architecture. For those use cases, data leaders determine requirements for new data-integration styles (often best addressed by data virtualization, specific data governance capabilities to fill risk gaps, and critical metadata / data catalog functionality to increase usability of data and create a base for better automation). They also consider the organization’s existing data-engineering capabilities to see what can be leveraged as part of their Active Data Architecture work and what metrics will be used to show the benefits of these changes.

A critical starting point will be engaging relevant resources, constituents, and teams to participate in planning and executing Active Data Architecture-related modernization work. This includes assessing skillsets and processes, and determining points of collaboration with existing teams and projects that can benefit from Active Data Architecture advances. Data leaders and their teams also must plan for how to measure and communicate the benefits (such as faster enhancements, improved performance, and great reliability), using quantitative metrics.

An Active Data Architecture is not something an organization can prepare and release all at once. Success with such evolutionary modernization, implemented in phases over a longer period, needs to borrow from the lessons learned from our data on successful business intelligence (BI) programs and initiatives. These best practices include:

  • Supporting only use cases that have clear business benefit and quantitatively measure the business value achieved
  • Starting with low-risk opportunities that will provide the “early wins” often critical to continuing program momentum
  • Using pilot projects to prove specific objectives; test approaches and methods; and validate assumptions about costs, benefits, and other relevant implementation scaling factors
  • Recruiting an executive advocate / sponsor
  • Promoting a data culture, which includes advocating data-driven decision making and improving overall data fluency through a formal data-literacy program
  • Communicating extensively with key constituents, including setting proper expectations

Phased delivery of an Active Data Architecture eliminates the risk of failure by trying to do too much all at once. Focusing on business value derived and obtaining early wins helps transform the Active Data Architecture concept into something more tangible and business-aligned, which makes it easier to obtain continued and perhaps additional or accelerated funding for future phases.

Data leaders need to start moving their organizations toward Active Data Architecture principles and capabilities. This includes the longer-term vision of a dynamic, self-organizing, and optimized architecture. Doing so does not have to be massively disruptive; they can start to introduce small and impactful automation actions (for example, automating decisions about the data-quality assurance aspect of governance, or making data placement and caching decisions based on workloads and performance metrics). The key is starting now to implement the most critical technology enhancements that will show immediate payback for key BI initiatives and also move the organization toward the overall vision of an Active Data Architecture.

You do not have permission to access this document. Make sure you are logged in and/or please contact Danielle with further questions.