What Is Data Governance?
Last Updated September 18, 2023
Data governance is taking on a front-and-center role among businesses as they grapple with and try to correct poor data management practices of the past.
This data management methodology is how an organization manages the data it employs, in terms of availability, usability, integrity and security. Developed and overseen by a governing body, it involves defining data management procedures and establishing a plan to ensure those procedures are followed.
The need to bring new rigor to data management partially springs from the explosion of data being generated. In 2012, 2.5 exabytes (2.5×1018) of data were created every day, according to IBM. That’s 2.5 billion gigabytes. Given this overwhelming volume, it’s not surprising that much of corporate data can be lost or become corrupt.
And there’s a significant price to pay. As Larry English, one of the earliest pioneers of data quality put it, “Process failure and information scrap and rework caused by defective information cost the United States alone $1.5 trillion or more.”
It’s leading businesses to devise holistic data management strategies that surmount the issues created by piecemeal approaches of the past. Typically, the tendency was to apply data quality and integration technologies locally or departmentally. That created silos of good data when more effective data management would traverse departments, applications, business units and divisions.
As companies move forward on the continuum from undisciplined to governed, they will have adopted a “think globally, act globally” perspective. They will have achieved a single view of the enterprise through master data management, gaining the benefits resulting from the integration of high quality data with business process management systems.
Data Governance Maturity Model
Better data governance (establishing, codifying and enforcing enterprise-wide best data management practices) is key for improving data infrastructure that was built for a different information era. The disjointed structure, with data held in disparate applications and multiple locations across the organization, results in lost or flawed data that hinders performance and drives up the cost of doing business.
What’s driving more holistic strategies today is the Data Governance Maturity Model. This helps businesses understand where their data management practices currently stand and indicates a path that can be taken to eventually evolve into a single, unified approach.
The model is based on four distinct stages for use of enterprise applications: Undisciplined, reactive, proactive and governed. Investments in both internal resources and third-party technology are necessary for each stage. Yet, as the organization evolves through each state, data becomes more universally consistent, accurate and reliable, making the rewards grow and the risks decrease.
Using the model, businesses can identify where they are on the continuum marked by technologies where data consolidation and integration commonly occur. Typically it begins with more limited areas like database marketing and moves through to more global areas like business process management integration. The continuum represented by the model is advanced, as businesses start with smaller projects to get more value from data (like database marketing) and then take on bigger projects.
A holistic data management strategy is what leading businesses aim to put in place, and the Data Governance Maturity Model is employed by many as a means to achieve this. As recounted earlier in this article, the model helps businesses understand where their data management practices stand today and indicates the path that can be taken to eventually transform it into a single, unified data management approach.
There are four distinct stages to the model. Systematically advancing effectively through them requires understanding the characteristics of each stage and what it takes to move forward.
Stage One: Undisciplined
In this stage, an organization has defined few rules and policies to set standards for data quality and integration. Executives aren’t likely to recognize the cost of data that’s been poorly managed.
A number of characteristics are common to companies in this stage. In the people category, for example, success rests on the competence of a few individuals. Management neither contributes nor buys in to data quality issues. On the technology side, undisciplined organizations do not perform data profiling, analysis or auditing, and cleansing and standardization are isolated occurrences. In terms of policies, there are no defined data quality processes, resources are not optimized and problems are addressed as they occur through manually based processes.
Organizations at this stage face very high risk levels, as data issues can force away customers and improper procedures can hamper productivity. Conversely, poor data quality returns few rewards. At some point, the price being paid for poor data is realized and effort is made to quantify the impact to spur change.
Advancing to the Reactive stage starts with establishing objectives for data governance, including the size and scope of governance efforts and what critical data assets are necessary. Also important are tech components that can handle data quality and data integration for cross-functional teams. It also is key to centralize, in a single repository, business rules for core data quality functions and use them across applications.
Stage Two: Reactive
As implied, companies at this stage tend to locate and manage data problems after the fact. Both ERP and CRM applications are used selectively, and while some employees understand the data quality issue, corporate management is largely unsupportive. Some 45 to 50% of organizations fall into this category.
Businesses at this stage typically have a group of database administrators or other employees who are responsible for data success. While data quality initiatives may benefit by individual contributions, there are no standard, non-siloed procedures. In terms of technology, tactical data quality tools are often available and are utilized by certain applications like CRM and ERP. However, data is not integrated across business units. While there are rules for data governance, the tendency remains reactive when it comes to data issues. Likewise, data management processes tend to respond to recent data issues.
This is another high-risk stage, given the lack of data integration and overall inconsistency of data. That translates into rewards that are limited, with returns delivered through individual processes and individuals given the overall lack of recognition of the benefits of data quality.
Advancing to the next stage (Proactive) is not easy. It requires creation of a new strategic vision to guide the processes for improving data that links to tangible business results. Moreover, it requires getting organizational buy-in, a challenge when business units have previously had significant autonomy over their applications and data structures.
The team uses best practices to establish cross-functional business rules for data integrity overseen by a data governance team that includes stakeholders and day-to-day data stewards.
Stage Three: Proactive
Organizations that reach this stage have a significant advantage in an optimal risk/reward ratio and an environment where data starts to become an asset that can drive informed decisions. Less than 10% of all companies have reached this stage.
This stage is characterized by management that understands and is committed to the role of data governance, understanding that data as a strategic asset. Data stewards maintain corporate data policies and procedures, and ensure ongoing monitoring maintains its integrity. Real time, not reactive activities, is the practice. Preventive rules and processes are in place as perspective shifts from problem correction to prevention. Risks are moderated because better, more reliable information is behind decision-making, and rewards are medium to high in light of improved data quality.
Moving to the next phase involves the solidification of the unified approach that is taking hold for all corporate information. Moving to the Governed stage requires the organization to assemble and integrate the pieces that have already been put into place. A framework emerges to organize the work of brand stewards, supported by businesses analysts and IT professionals. With data now robust and reliable enough to for high-end process management, the foundation for Business Process Management integration is now in place.
Stage Four: Governed
At this stage, data quality, integration and synchronization are integral to business processes under a unified data governance strategy.
In this environment, the CEO is directly behind the strategy given its executive level sponsorship. Employees are involved in data strategy and delivery, and zero defect policies rule. Data quality and data integration tools are standardized while data is continually inspected, with deviations immediately resolved. New initiatives are weighed carefully against their potential effect on the existing data structure. Automated policies ensure data’s consistency, accuracy and reliability enterprise-wide.
Risk is low given the tight controls over master data, organization-wide that ensures high quality information about customers, prospects, inventory and products. Rewards are also high, given the improvement of data-led insights into the business that increase management’s decision-making confidence.
The company that reaches this stage has effected a major culture change, where managing data is less a tactical challenge and more the basis of a sophisticated data strategy and framework that helps drive the business – and is a significant competitive differentiator.