Getting and Keeping Your Data Right
Enterprise Data Asset Management
Published: May 1, 2008
Organizations often do not realize just how poor and out of control their data really is because they hold tight to the false belief that they can make the data “right” by simply making it look right.
As our world has transitioned from an industrial age to an information age, data has become a very valuable resource since it is the raw material used to produce information. In this perspective, data should be as important to a business organization as any traditional asset, probably even more. In spite of this obvious fact, as well as all the recent hype regarding enterprise data asset management, most organizations continue to be data challenged as their data grows further out of control. This begs the question, if data is such an important enterprise asset, why is it not managed as one? Do business organizations really understand data; after all, you cannot manage what you do not understand.
Data is simply a representation of the real world business organization. A representation “stands in for” whatever it represents; it depicts or is a likeness of something. It serves to capture the essence and account for the known or inferred properties of the real world thing. Everything in a typical business organization today is based on this data representation, from recording and operations to predicting, planning and reporting. Fundamental to all these activities is decision making. Rarely in today’s world are business decisions made on actual observations, but rather using this data representation. The ability to make accurate, timely decisions is essential to the survival of any organization and requires the truth: integrated, clean, reliable, accurate, managed data. This gives an organization the strategic advantage to compete and win; it can actually make or break an organization. Getting the data right and keeping it right is what guarantees the truth. Enterprise data asset management is all about getting and keeping the data right.
Data as an Enterprise Asset
An asset is defined as any item of economic value owned by an entity that can be measured and tracked. Assets can be physical such as cash, inventory, or property; or intangible such as patents, copyrights or trademarks. Data clearly has an economic value to an organization because if it were destroyed, lost or stolen, there would be a significant cost. In most business organizations today, data has become as critical in the creation of products and services as equipment and materials; data can sometimes be a product itself.
Data assets differ from the typical business assets in several ways. Data is not depleted when used; it can actually grow as it moves through an organization. The physical location of the data asset is not as important because technology provides a virtual environment. Regulations, laws, and taxes typically assigned to other types of assets present much more of a challenge to assign to data assets. This is mainly due to the elusive nature of the value of data, as its value depends on situations, content, human competencies, and time. Downsizing in regards to physical assets is often seen as a good cost-cutting method, but that is rarely the case for data assets. In spite of these differences, data assets still need to be tracked, measured and accounted for in much the same manner as the traditional physical assets. What is not tracked and measured cannot be managed.
The majority of what’s written today about how to treat data as an enterprise asset gets twisted into a tools or technology sales pitch. Many of these methods start with the existing broken data, overlooking the most essential step – getting the data right in the first place. If the data was never right in the first place, then all the asset management methodologies in the world cannot make it right. Since data is used as a representation of the business organization in order to track, manage, and predict it, an accurate representation is absolutely necessary in order to begin to treat data as an important enterprise asset. The entire organization is dependent on a correct representation, as every single function within the organization, including enterprise data asset management, hinges on this.
The Foundation – The Enterprise Data Blueprint
If it is critical to get the data “right” in order to manage it as a valuable enterprise asset, then the first step in any enterprise data asset management initiative is to have a clear, correct understanding or picture of the business organization – a data blueprint representing the real world organization. Without this data blueprint, how would you ever know if your data is right and by what criteria would you judge? It is very similar to the architectural blueprint used to build and maintain any physical structure or building, where everything is evaluated against this plan. The data blueprint or map of the organization needs to represent one single version of the truth, not any one specific view or functional usage of data. It must represent the business organization as a whole, its things and events. Most technology projects focus on a specific business functional use of the data. The data blueprint, unlike a logical data model that typically represents a specific business usage of the data, represents the business organization, and these are very different concepts. Once this data blueprint, also known as an enterprise data model is complete, then everything else needed to manage data as an enterprise asset can be accomplished successfully.
While many organizations recognize the need for their data to be right, they attempt to “clean,” or fix it without this data blueprint. How would they ever know if the fix correctly represents their real world organization? Unfortunately, there is an unconscious consensus that if the data “looks” right, then it must be right. That is one of the biggest data fallacies today and underlies many of our data challenges. Data quality statistics account for only the data that is obviously or proven incorrect. As long as the data does not appear to be off, it is assumed correct. Think about it, if you already know the correct answers, you would not need to ask the question. So when questions are asked of the data, most assume the answers are correct, as long as they do not look obviously incorrect; blindly, many believe their data merely on an assumption of it being correct, not ever really knowing the truth. Therefore, within what is assumed as good quality data, there is likelihood of error. Attempting data repair without a data blueprint often moves the obviously incorrect data to this assumed correct state. This data is then used to plan, operate, predict and steer our business organizations.
Instead of creating a data blueprint, often an organization chooses to “purchase” an enterprise blueprint such as an enterprise package or a generic industry data model. These can represent an organization theoretically, but it is still essential to develop a data blueprint of the actual business organization. How would you ever know if the purchased picture is a correct representation of the actual business organization? If you had the actual data blueprint of the organization, then it would not matter if an enterprise package forces a black box or generic implementation because the differences could be identified, understood, documented, mapped and managed. This includes all of the deviations from the truth that the package solution presents. The same thing holds true for a generic industry data model. The organization’s data blueprint can be measured or compared to an industry view in order to understand and manage the differences. Just like people, organizations are all uniquely different, and it is this difference that can give a competitive edge. A data blueprint of the real world business organization is absolutely essential in order to begin to manage data as a critical enterprise asset.
Enterprise Data Asset Management
If a business organization actually viewed their data assets as they did their traditional assets, there would be little doubt that the data would get managed properly. For example, a typical delivery service business understands their delivery vehicles as critical assets. They know the importance of having each correctly configured and appropriately maintained. It is essential that they uniquely identity each vehicle and keep an accurate inventory at all times. If their fleet were initially configured inaccurately or tracked and maintained improperly, then the business operations would be compromised and inevitably fail. For example, a business specializing in appliance delivery with vehicles not tall enough to accommodate refrigerators or improperly maintained will result in missed deliveries and angry customers. These same principles hold true for the data assets.
One of the biggest issues facing enterprise data asset management is maintaining the quality of the data asset. A data quality program for the ongoing maintenance of the data asset needs to be established. Fundamental to a data quality program is the knowledge of both the nature of data and its quality challenges. This is where the enterprise data blueprint plays a vital roll, in both understanding and prioritizing the data for quality management. Ownership can provide the governance as well as the accountability for the data quality. Data quality expectations, measurements, methods and audits need to be established. Tied to data quality is data integration. Here again, the enterprise data blueprint provides the fundamental understanding of how the data is integrated, as it is the one version of the truth and will assist in the resolution of redundant and disparate data challenges.
The simplest of all management techniques is prevention. Today, more than ever, organizations need to address enterprise data asset management from a point of prevention, eliminating anything that absorbs system and staff resources while providing little long term value. This should include all of the resources that continually band-aid the data quality symptoms, like giving an aspirin to a person with a brain tumor. The typical approach to data used in many organizations today is very reactionary, chasing the symptoms because most organizations feel they can not afford the time and resources to get their data right in the first place, as urgency always trumps strategic. All data problems severely handicap business plans by draining productivity and capital as well as inhibiting agility and growth. The irony is that organizations in today’s world simply cannot NOT afford to get their data right because of the very critical and strategic nature of this asset.
The first challenge many organizations face in regard to managing data as a valuable enterprise asset is to first recognize that their data problems severely hinder their organization. Nearly all organizations admit their data is riddled with issues, but believe it is nothing they can’t handle. All they have to do is just fix whatever does not look right with the magic tools. Simply do whatever it takes to make it look right. Most do not see that there is something seriously wrong with this approach. These organizations do not realize just how poor and out of control their data really is because they hold tight to the false belief that they can make the data “right” by simply making it look right.
Bad quality data is but a symptom of the problem and not the problem itself. This scenario is similar to the typical alcoholic who does not think he has a problem because he still has a job or owns a home, and thus thinks he is in control. It is interesting that the inability to recognize an alcoholic problem is a big part of the problem in itself, exactly like our data problems. We could say many organizations are dataholics. Add to that the typical codependent behavior of the individuals responsible for the data – how would it make them look to recognize that there are serious data problems? It could actually be a career-ending move to call the baby ugly.
Once there is recognition of the severity of the data problems, a business-driven holistic approach, mandated from the executive level, is needed to establish data as an important enterprise asset. This often requires a cultural transformation because typically the information technology arm of the organization drives the data strategy, as most organizations assume data to be a technology discipline. Data is clearly a strategic business asset and needs to be owned and managed by the business, not technology. When data is driven by technology, much of the organization’s critical data ends up in departmental silos, with each group using different technology approaches, resulting in the lack of a holistic view of the data. It requires bold executive leadership and financial support to establish data as a critical enterprise asset. This also takes experienced data professionals that understand the nature of data and have the expertise to properly manage it. An enterprise data asset management initiative cannot be successful if it is driven from the bottom up, chasing the data symptoms with tools or technology.
Enterprise data asset management is a strategic business initiative. As with most strategic initiatives, it is often the first to be cut in the wake of economic risks or challenges because it is considered non-critical activity. It is very difficult to account for the return on investment that fails to directly link to the bottom-line contributions of the business organization. It has been a challenge to understand the nature of data enough to establish and attach actual value as very few organizations have successfully incorporated this understanding. The lack of a strategic approach to enterprise data asset management is also fueled by the short-term thinking within most organizations today as they compete in the new world economy. Ironically, managing data strategically as a valuable enterprise asset will assure the survival of a business organization in the future. The organizations that figure this out first will be leap years ahead of the rest.
Recent articles by Noreen Kendle
Noreen Kendle -
Noreen is currently the Senior manager of Data Architecture for the Home Depot. With over 25 years of experience in information technology, primarily within data architecture, she has worked within the communication, financial, manufacturing, health, non-profit, retail, graphics, and travel industries. Noreen has extensive industry experience leading the development and implementation of data architectural initiatives including: enterprise modelling methodology, master data, enterprise data management, data ownership, enterprise data integration, data quality, enterprise business intelligence, metadata, data governance, and physical design. She can be reached at email@example.com or 678-768-6591.