Building Trust in Data Through Governance

Gary Gilbert,
Software Architect

In his 2017 book “The Inevitable,” Kevin Kelly states that technology is in a perpetual state of “becoming.” That is, technology is never finished or done - it is just a series of changes. I see this concept of “becoming” in the current data landscape. New tools and services, new patterns of consumption, and even new architectures arrive on the data scene on a seemingly daily basis. Given the breadth and pace of change, data governance has never been more important. In fact, because of the constant flux of “becoming,” Kelly argues that processes like those put in place by a strong data governance program have become more important than products. In this post, we will cover some of the key characteristics a data governance program must have to succeed in this time of becoming.

Traditionally, data governance has been a defensive mechanism focused, for the most part, on legal and regulatory compliance. While those defensive responsibilities are still there, with the rate of change and distributed nature of modern data architectures, a data governance program must now also play an offensive role. That notion requires a mindset shift away from governance as the enforcement of rules, and instead toward governance as a set of processes and tools.

A modern data governance program must include a flexible organizational structure, capable of providing both strategic and tactical guidance. Typically, the data governance organizational structure includes, amongst other roles, a data governance council and a director-level resource. The governance council is composed of people from across the organization, including product owners from both data-producing and consuming systems, and technical resources. The role of the governance council is to set the priority across data initiatives, develop policy (e.g., a data retention policy), and collaborate with technical staff to inform data models and contracts. The director plays a more tactical role in leading implementation of policies defined by the council, and is ultimately responsible for the delivery of data initiatives.

With the organizational structure in place, the offensive responsibilities of a governance program boil down to one thing: trust. Simply put, if consumers don’t trust the data, how can they be expected to use it? Therefore, a modern data governance program must be laser-focused on tools and processes that enhance the trustworthiness of data. Quantifying trust is difficult, but trust can be grounded in the ability to answer the following questions.

  • What does this data represent?
  • Where does this data come from and how has it changed?
  • Where can this data be found?
  • How has the accuracy of this data been determined?
Put another way using industry terminology, does the data governance program have processes and tools that support data definition, data discovery, data lineage, and data validation?

While it is possible to use a piecemeal approach to provide these key features, many data governance programs leverage a data catalog for, at a minimum, data definition and discovery as well as lineage and validation in more advanced programs. There are many data catalog options on the market, and an in-depth review of data catalog solutions is beyond the scope of this blog post; suffice it to say that a data catalog serves as a record of all data in an organization and can be used as the central tool for building trust in data through definition, discoverability, lineage, and validation.

Trust in data can also be enhanced through transparency. The modern data governance program builds transparency by ensuring that tools and processes are not only well-documented but well-understood across the organization. If data is an asset then it should be considered the responsibility of everyone in the organization... but that means people must be aware of the tools and processes needed to fulfill their role. A modern data governance program should include a communication plan that broadcasts information about key initiatives and policies, as well as a training plan so that employees have the skills and understanding needed to take advantage of the tools and data.

Hopefully, this post was able to help you rethink or reinforce your current thinking on governance in this time of data “becoming.” My own experiences with data projects have led me to the conclusion that data initiatives without a strong governance program may succeed, but they do not thrive or scale very easily. In these cases, I often see a solid technical infrastructure but issues with usage because the data is not discoverable, or there are issues with data accuracy as new datasets are onboarded due to a lack of validation and lineage tracking. In both situations, the customer is left to reevaluate and retrofit their governance program while still operating their data infrastructure - not an easy task. Unicon’s Data Strategy Evaluation includes a review of your current state, desired future state, analysis of what’s missing, and a clear roadmap on how to get you there. Included in this evaluation is the focus on process and governance. Helping your organization with the overarching data strategy that includes focus on governance ensures you have both an offensive and defensive game plan for Data Success.
Gary Gilbert

Gary Gilbert

Software Architect
Gary Gilbert is a Software Architect at Unicon where he provides technical leadership to the integrations and learning analytics practice. He has 18+ years of experience designing and developing learning systems for clients ranging from small community colleges to global publishers.

Gary has been involved in numerous open-source software projects including Sakai, uPortal, and Moodle as well as open standards efforts such as IMS Learning Tool Interoperability. Gary has been involved with IMS LTI since its inception. Over the last 10 years, Gary has led dozens of IMS LTI based integration projects and has experience developing integrations with every major learning management system. Additionally, Gary has led the development of several integration platforms that support open standards-based integrations as well as custom integrations and managed the entire integration lifecycle from implementation to client on-boarding and support. Gary is particularly interested in the intersection of open learning technology standards and open educational resources. Gary specializes in Java, PHP, and Javascript. He is experienced with most mainstream application frameworks with a focus on the Spring Framework.