Let’s take a look at what it is, why it’s important, and how your enterprise can leverage it to reach your fullest potential.
What is Data Governance?
Data governance is defined as the specification of decision rights and an accountability framework to ensure the appropriate behavior in the valuation, creation, consumption, and control of data and analytics. It’s an umbrella term that covers distinct tasks, enabling a business to preserve, sustain, and extract value from data.
Why enterprises need Data Governance
The goals and advantages of data governance are too many to name, but some of the most important are:
- Make confident, data-driven decisions
- Decrease the risk of regulatory fines
- Improve data security and minimize the risk of a breach
- Maximize your data’s income generation potential
- Designate accountability for information quality
- Enable better planning by supervisory staff
- Minimize or eliminate re-work
- Optimize staff effectiveness
- Establish process performance baselines to enable improvement efforts
These goals reflect the pillars of DAMA International’s famous data management wheel, as seen in their DAMA-DMBOK (Data Management Body of Knowledge) framework. We’ll cover it in more detail below.
What goes into Data Governance? Pillars and prospects
There are a few ways we can visualize the inner workings of data governance: the wheel and the house hierarchy. First, let’s take a look at the wheel.
The wheel represents all the pillars of data governance, upon which the success of the entire system rests. The implication being that if one pillar is neglected, it could jeopardize the entirety of an enterprise’s analytics efforts.
To put things another way, we can visualize the same end-effect hierarchically (below).
In this image, you can see that all the processes are interdependent pillars, which together comprise the umbrella concept of data governance. Ideally, each element receives equal weight from the start of implementation and is developed in parallel with the others.
How important is Data Quality in the scheme of Data Governance?
Of course, without quality data, there can be no quality decisions. But while most enterprises are aware of that most obvious pillar – data quality management – unless you address data security, that data could be susceptible to security breaches. For enterprises, this poses a potential threat on multiple levels.
Other pillars ensure that data is properly organized, to make it easy to retrieve when it’s needed and maximize its value (data operations management). Still others ensure that data from multiple sources gets appropriately integrated (data development/data integration).
Usually, we see data quality as the critical point of data governance, but in reality, it’s only one slice of the pie. The data governance pie involves other, just as critical aspects of dealing with data, like data architecture, development, security, and the rest of the pillars we see in the images above.
New streams of analytical systems: Data Lakes and Data Warehouses
In the past decade, new streams of analytical systems have emerged under the term “Big Data.” Existing approaches could no longer handle the significantly larger volumes, velocities, and varieties of unstructured data available. Therefore, new architectures oriented toward Big Data rapidly gained momentum, resulting in an architecture known as “data lakes.” While data warehouses mostly store processed, structured data, data lakes store vast amounts of semi-structured and unstructured data. Unstructured data is omnipresent today, not only in enterprise systems but everywhere online.
Document and content management
While also present in data warehouses and business intelligence, document and content management has a more significant role in data lakes.
Document and content management as a part of the data governance processes helps ensure the availability of those data and their efficient retrieval while helping to ensure business continuity through best practices and lower operating costs.
Finally, metadata management in the era of data lakes has become more crucial than ever. The established metadata processes that hold the stored data have a significant role in creating and maintaining the proper data lake by having notably heterogeneous data.
Data governance in Data Lakes and Warehouses
With ever-increasing amounts of data to process and so many key decisions depending upon their accuracy, failing to implement data governance processes on time can lead to chaotic messes of jumbled information – and ultimately a complete lack of trust in the system. And as that system grows, the problems multiply.
Warehouses, as opposed to lakes, are better equipped to handle these problems as their data volume and variety are typically smaller. New data gets loaded into warehouses in larger intervals – typically daily. On the other hand, data lakes deal with heterogeneous data that arrives in bulk, and several times a day. Timely implementation of data governance processes will prevent your data lakes from becoming data swamps, where mountains of useless data bury the helpful information — rendering it useless and inaccessible.
Where is your enterprise on the Data Governance maturity ladder?
Every data governance system eventually reaches some level of maturity. Typically, those levels grow from fragmented systems, and the data governance is driven more by the technical staff and less by the business stakeholders. As the system matures, the processes are increasingly business-driven and give a more holistic overview of the enterprises’ needs and usages.
The IBM Data Governance Maturity Model
The following image shows the maturity model proposed by IBM. Below, we’ll break down what each level means for your enterprise.
Level 1: Initial
The trajectory begins with little to no awareness of how important data is, and no standards established for managing it. The data is typically stored in silos, with an ad-hoc, informal approach to data management, with systems being built request by request without deadlines or processes in place.
At this level, system architects play an essential part in studying the data, information flow, the enterprise’s needs, and detecting any data-related vulnerabilities. A data management plan has to be established and approved by the business and IT stakeholders to mature beyond this level.
Level 2: Managed
Enterprises reach this level when they understand the importance of stored data, and how it can be used to the benefit of the organization. With data now considered a valuable asset, the enterprise can implement data management tools and governance processes, and establish documentation guidelines and regulations.
Level 3: Defined
The guidelines and regulations now get integrated into the processes that enable stored data to be used at the enterprise level. As a result, the technology used for data management is more tailored to the enterprise’s needs, while the data governance processes and practices are present at all levels of the organization. By this point, risk assessments for data quality and management are usually part of standard operating procedures.
Level 4: Quantitatively managed
In this stage, all ongoing and newly established projects utilize the data governance guidelines and principles. All data models come with thorough documentation and are available to users in the organization. Projects at this level get the data quality goals and KPIs with people responsible for them, and continuous measurement of their performance in given objectives.
Level 5: Optimizing
By the time an enterprise enters the optimizing level, its well-established data governance processes have rendered its data much more manageable, resulting in greatly reduced data management costs. All data projects are continuously monitored to determine ROI and compliance with the data governance rules.
Leveling up your data governance processes
Each enterprise has a unique path and timeline to reach a particular maturity level. No matter what your goal is, your organization needs to dig deep into its assets, assess its current level, and create a strategy for implementing data governance processes.
The importance of Data Governance for your organization
Data governance empowers enterprises to propel their business objectives forward with agile, confident, and data-driven decision-making.
Its processes enable the gathering, organization, and utilization of critical data assets, ensuring that the data in your analytical systems can be trusted and used as a single source of truth — and guaranteeing that your business intelligence is in fact, intelligent.
The benefits of data governance are crystal clear:
- Data consistency, quality, and accuracy
- Improves data confidence and trust
- Improves business planning
- Maximizes profits
Data governance is a process that evolves over time – not a one-and-done task. If you’re considering implementing data governance in your enterprise, our expert team is here to help you do it right.
Author: Tomislav Hlupić, Senior Consultant