What is data governance and why does it matter?

DATA GOVERNANCE IS BECOMING INCREASINGLY IMPORTANT AS A KEY BUSINESS PRIORITY.

Matt Sanson, President of The Virtual Forge North America, delves into the crucial role of data governance in today's increasingly complex landscape.

What is data governance?

Data Governance, simply put, is the pursuit of increased quality, understanding, and control of the data within an organization. It helps organizations to capture and apply business expertise and technical expertise regarding data, so that they can more consistently create and distribute trustworthy information.

It is successful when it reliably connects high quality information to data consumers who can use it to increase the performance of the business.

A Data Governance program encompasses people, processes, and, ultimately, data governance technology that are applied to manage data and its uses throughout an organization. The specific components of the program and activities may vary from program to program but typically include the following:

• Governance Framework - this is the overarching program design to define the program components, program goals, and the organizational design behind it;
• Metadata Collection - tools and processes to document the data and its use within the organization. This may include Business Glossary (to capture business terminology associated with reporting), a Data Dictionary (to capture details about the data as it resides within the organization’s systems) and a Data Catalog that maps between the physical data and the business terms;
• Stewardship Model - the means by which people are assigned accountabilities for data and data governance activities. Business Stewards and Technical Stewards are responsible for collecting and maintaining metadata and providing approvals for governance actions.
• Use Case Inventory - this is a log of the most important uses of data within the business along with a detailed estimate of the business impact of each. This can be used to create a backlog of future data projects where the anticipated ROI of each can be used for prioritization purposes.
• Data Lifecycle & Policies - mapping and managing the ways in which data is created, used, stored, archived, and disposed. This allows the organization to get their arms around the supply chain of data within the organization and is a foundational step for data security and data privacy management.
• Data Quality Management - these processes allow organizations to define what good data looks like in the context of their operation, to monitor the quality of their data based on those definitions, and in some cases to remediate the quality of data where it is found to be lacking.

The activities above all help to create an effective data governance program. However, data governance is not something that can be completed and checked off a list. It is an ongoing process that must be continuously operated and enriched to produce results.

Data and the way it is used is always changing, and so must the framework applied to understand and curate it. Much like diet and exercise, data governance can be a difficult regimen to maintain as it is hard work, and can lack the satisfaction of quick results. With persistent and disciplined use of the process, a wide range of benefits can be achieved. Data governance creates data fitness.

Why does data governance matter?

Data Governance processes aim to understand and curate data so that the people within the business can find insights and advantages, eliminate risks, and become efficient custodians of the information they create, store, and use.

The importance of Data Governance to an individual organization has a lot to do with their business strategy and their unique data footprint.

Do they operate within a regulated industry or location that has specific legal requirements that drive a regulatory compliance and risk-focused governance approach? Does the company operate in a financial marketplace where information advantage in sub-second time frames can be the difference between success and failure, which would favor an agile approach that prioritizes integration and enrichment? What are the goals of company leadership in relation to governance, and how do the capabilities of the people who would operate the Governance process match up with the tools available to execute those tasks?

Whatever the most important result for the organization, anyone who uses data as a critical part of their operations stands to benefit from the efficiencies that proper governance creates.

This can be counterintuitive because governance activities are often seen as expensive, which they absolutely can be. Policies, data management and quality tools, metadata repositories, and the time of very talented people to make them all go can bring with them substantial budgets.

The return on those investments, while substantial, can be more difficult to quantify. The benefits are diffused throughout the organization, and take many forms, so it is critical when creating a governance strategy to quantify the expected business benefits of any projects to be pursued at the outset, and to create measurements that will clearly identify where ROI is generated.

It is only with this context that the wider business, including key leadership roles, will understand the value of the governance process as a tool to accelerate the business. That framing will help to ensure the Data Governance effort is sticky even when challenges arise, as the value will be widely understood and measurements will be in place to monitor progress.

What makes data governance challenging?

Within a company, data can be an ethereal thing - people and processes both need it to operate properly, and produce it when they do. It’s an IT thing, it’s a business thing, it’s strategic insights, and it’s useless noise. Data flows everywhere in a business, and it does not follow the org chart as it goes. This inherently makes “ownership” of data a challenge, because everyone has an interest, and a unique lens with which they use to look at it. Each group may feel some ownership for the data that they use, but if “everyone is in charge” of something, the reality is that nobody is accountable.

As data is created, modified, consumed, and piped to different places, it can be changed (for the good or the bad), it can be labeled different things, and mixed and combined to create new data products.

Without centralized governance this will lead to differences in the data, and these differences can prevent the business from seeing their critical insights in a unified and agreed way. Inconsistent results and different understandings of the data are created, and with it mistrust of the information across the groups that consume it. Data consumers that don’t agree on what the data says will create factions within a business, and a lot of wasted time. These groups feel compelled to take control over their own data and produce it the way that they trust it. Ultimately this duplicates data operations activities across the company, and sharply reduces transparency and efficiency.

In the absence of formal governance, people create their own systems for determining what information can be trusted, and how to source it. These processes can become ingrained in the way that a company operates, and attempting to change those systems can be met with a great deal of resistance.

People become attached to these processes as they deliver a critical service to them, which allows them to overcome an organizational gap, but ultimately using a workaround vs. a sustainable solution. In this scenario, data consumers trust the data that comes from the PERSON who they believe has the best understanding of it and who can pull out the insights. Data Governance simply aims to replace that trust in a PERSON with trust in a PROCESS that can leverage the skills and knowledge of many different stakeholders to produce the desired results. This process should be managed above the individual operating groups, so that the organizational goal of data consistency can be baked into the process, while stewards representing parts of the business or individual data sets can participate directly in the process to champion the unique requirements of their individual groups.

Another challenge with Data Governance is the way the benefits are delivered, as the efforts are front-loaded and the results require time. Starting a governance program is like starting a new business unit, and requires time and effort to build the baseline capabilities and practices, to upskill teams to take advantage of these new capabilities, and to allow adoption and maturation of the process to take place. The company must take the time to not only wrangle the current data and understand how it could be effectively applied by the business, but also to execute on those targeted use cases so that the envisioned value can be achieved. Governance on its own is at best just cost and risk reduction, but trusted data that puts precise insights in the hands of the right people can transform businesses entirely. It requires the vision and patience to reach this stage, and the organizational support behind it.

The key ingredient

Let’s face it, Data Governance sounds difficult.

We know already that the effort can be time consuming to put into place, and that the benefits are delayed relative to those efforts. The people who need to spend time on the process as stewards are the data and business process experts, and their time is always in high demand. Even the term has negative associations, because it sounds slow and restrictive. Because of this, governance can have a bad rap, and there is important work to be done to educate and communicate with stakeholders in the business to ensure they understand what is being done, and where they can expect to benefit from the process.

The destination where both the business and IT understand the information at their disposal, where definitions of metrics and data use are agreed and consistent throughout the business, and where insights are rapidly discovered is a worthy objective and one that pays the business back many times over. However, to get there, what is required is leadership. An individual, or a collective within the organization must be given the resourcing and authority, and must possess the confidence and understanding to take the business on this journey.

The data leader

The Data Leader is a difficult role to fill. Ideally this individual must be able to bridge the gaps that exist between IT and the business and communicate clearly with both. They must have a deep understanding of data governance and strategy, and the ability to communicate it to all levels of the organization and influence at all levels of the organization. They must ensure the program, the strategy, and its actions are always in alignment. They must be able to learn the nuances of how the business operates with data today, and also have the fortitude to introduce and champion changes that have the potential to make it better. Yes, we’re looking for a Unicorn.

In large organizations, Data Leaders often take the form of a CDO, but many organizations don’t have the scale or the budgets to create a dedicated org to support these efforts. This means that in many cases, the teams, and at times even the leadership, take on the program as an additional responsibility vs. a sole focus. I would always advise that the data leader should be a full time role, but it is a reality that many organizations just have to do the best they can with what they have.

So to select the right person from existing staff or from outside, what is most important? Historically the CDO role has been thought of as a technical position, where the critical skill sets have to do with the technical aspects of data and data management, and the associated applications that are used to execute those processes. While technical understanding is definitely helpful for a data leader, there has been a gradual shift over time to this role that favors a strong business and analytical background over pure tech.

The reason for this is that over time businesses have learned that the business-led Governance efforts tend to work the best. Starting the program with a precise view on business benefits that can be attained ensures that the organization can be targeted to govern and improve the data required to achieve those benefits. This helps to keep the program at a reasonable and efficient size as it gets off the ground and produces its first successful outputs.

Getting the requirements right and the data product output correct so that the business can make use of them takes a lot of facilitation and collaboration with the business stakeholders. The governance artifacts that are most difficult and time consuming to produce typically involve the sourcing and mapping of information that exists within the heads of business users rather than in the technical systems.

This gives an advantage to a data leader with a business mind so that they can facilitate those communications and remain focused on how to make the business better with data. If this is done properly, and business definitions and requirements are effectively sourced, the project elements that are more IT related can be handed off to a dedicated IT team, who will have an improved ability to execute when provided with the detailed context the governance process creates.

New normal

Data Governance may continue with its bad reputation far into the future, as it is hard work, requires investment, and generates process change which may be uncomfortable for some stakeholders. However, these processes are here to stay, and organizations who can overcome the startup barriers will be well positioned to meet the data challenges of the future.

Every business should strive to create trust in the data that it uses, because without that trust, the information loses its decision making value, and can become a risk vs. an asset. It is highly unlikely that the quantity of data used by businesses, or the strategic criticality of that data will be reduced in the future. So with greater data velocity, variety, and consumption as realities, data governance and management programs will be crucial to assist organizations in making sense of what they have, and enhancing their ability to quickly, effectively, and safely apply that data to execute their strategy and find competitive advantages.

How can The Virtual Forge help?

Our global team of data experts offer custom solutions and consultations around data architecture and data governance.

Our goal is to assist anyone in overcoming their data challenges in a rapidly evolving and increasingly complex world. To gain a competitive advantage, every aspect must be carefully managed. If you're facing a challenging data issue affecting your business, contact us for a free data governance strategy session. Our specialists are always ready to have a productive discussion and help you find a solution.

With over a decade of experience in industries such as automotive, pharmaceutical, retail, finance, and more, our team understands the complexities of each sector and how to navigate them. Our consultative approach enables us to better understand your organization's needs and goals to ensure success.