Register now for - "Analytics for Power Users: Practices and Products You Need to Know" - Wednesday, December 13, 11:30 a.m. Eastern Time.

The Path to Modern Data Governance

Data governance problems

Natural tensions exist between the needs of self-service data consumers, the values of agile practices, and the traditional practices of data governance. I first wrote about The Next Generation of Data Governance in 2017 and I still see companies struggling with the conflicts today. Complexity and rigor of traditional governance conflict with agility and speed of agile analytics. Decision rights and policy enforcement inhibit the autonomy that is fundamental to self-service. Self-service analysis is often iterative, exploratory, discovery oriented, and not highly compatible with rigorous change management processes. These are but a few examples of the tensions between two data-related areas that are culturally and objectively misaligned. 

We need to modernize data governance to be less process focused and more people oriented. The reality is that we don’t govern data. We govern what people do when working with data. And it is in the human dimension where much of modernization must occur. We need to change governance practices related to policies, complexity, decisions, and rigor—shifting data governance from a controlling organization to a service organization. Most data-driven organizations agree with these goals. The real challenge is how to make it happen. In this article, I offer a framework and some guidelines to help create a roadmap to modern data governance.

A Modern Data Governance Framework

The data governance framework shown in figure 1 makes it clear that governance modernization isn’t quick and easy. It is a journey, not an event. There are many parts to data governance and planning is critical to modernization. Among the parts:

  • Goals describe why we govern data. Without goals, governance lacks the focus needed to deliver value and align with business objectives.

  • Methods describe how we govern data. Policies and the guides, guardrails, and gates needed for prevention, intervention, and enforcement are the primary methods. A code of conduct to guide ethical judgments is an important governance method when addressing data ethics. Curating and coaching are supporting methods that help people to comply with policies and to make good judgments.

  • People are at the heart of data governance. The reality is that we don’t govern the data. We govern what people do when working with data. The full range of people – sponsors, owners, stewards, curators, coaches, consumers, and data stakeholders must be considered. The scope of stakeholders can be quite broad including individuals that may not be immediately apparent – internal auditors and risk managers, for example.

  • Processes are the series of actions taken to achieve specific results – to resolve issues, to coordinate changes, to assure data quality, to catalog datasets, and to measure and monitor data governance impact.

  • Technology fills many roles in data management – data ingestion, data cataloging, data preparation, data analysis, and data pipeline management. Each of these technologies offers some degree of data governance features and functions.

  • Culture establishes the behavioral norms to create an environment where data governance, agile projects, and self-service data analysis and reporting all work together without conflict. 

Figure 1. A Modern Data Governance Framework

The framework has six layers, and at the most granular level it consists of 39 components. Each of these components needs to be addressed as part of data governance modernization. Use the framework as a structure to analyze modernization needs, and to prioritize and sequence modernization projects and activities.

The Modernization Roadmap

Take time to analyze your current state and your modernization objectives for each of the framework components to determine what specific things you need to do to evolve your data governance program. Figure 2 shows an example of modernization activities derived from framework-based analysis. The list on the right shows many of the common activities of data governance modernization. Note the color coding that connects the activities to layers in the framework, and therefore to the planning questions from which the activities are derived.

Figure 2. Data Governance Modernization Activities

It is worth noting that the longest list of activities is the people list. This is typical, as having all of the right people, engaged in the right ways, is critical to data governance success. The processes and methods lists are tied for 2nd longest. People, processes, and methods are at the center of effective data governance.

The example shown in figure 3 illustrates the idea that we have selected a subset of the activities – not all of them – for initial planning. (The color coding here is different, mapping activities to projects.) To make modernization manageable and practical, it is important to make conscious decisions about what NOT to do. The selected activities are organized based on affinity – they seem to fit together and make sense as a project. They are also organized and based on dependencies – what makes sense to do in what sequence. Note here that the activities in a single project don’t necessarily all come from the same layer of the framework. The bottom sequence in green, for example, includes two activities from the culture layer, one from the methods layer, and one from the people layer.

Figure 3. Modernization Projects

Once projects are defined as sequences of activities, the next step of road mapping is to plot them on a timeline. (See figure 4.) Note that in addition to the four projects identified in figure 3, a 5th has been added. When plotting projects on a timeline we realized that engaging the right people early in the journey is important, so added project #2. This creates a high-level roadmap of activities and timing for the governance modernization journey.

Figure 4. Data Governance Modernization Roadmap


The roadmap, of course, should be a living document – a plan that changes as needed. This example is plotted as a year-quarter-month timeline. Ideally, the roadmap is revisited once each quarter to identify and adjust for changes in needs, priorities, schedules, or other factors. 

The Modernization Imperative

So many aspects of data management have changed in recent years – big data, data lakes, self-service, cloud, agile, DataOps, etc. that we can’t continue to govern using turn-of-the-century methods. To succeed as a data-driven organization you must modernize data governance practices. The changes are many: from data focused to people focused, from control to management, from gates to guides and guardrails, and much more. Modernization is a journey that will involve a series of projects and affect many people. My recommendation: start with planning and road mapping to avoid pain and false starts.

Dave Wells

Dave Wells is an advisory consultant, educator, and industry analyst dedicated to building meaningful connections throughout the path from data to business value. He works at the intersection of information...

More About Dave Wells