There are always new terms being introduced in the data world. While newfangled terms like “data mesh” and “data fabric” require lengthy descriptions, the term “critical data element” or CDE is easier to explain. But… as the name implies, CDEs are critical – not only to data management, data governance, and data quality success, but to the success of the enterprise. In this article, I quickly describe what a CDE is and why CDEs are essential to your business.
Before I get started, it is important to note that individual data elements are subsets of data resources that may house hundreds, if not thousands, of data elements. Best practice has demonstrated that organizations are better off focusing on subsets of data rather than trying to improve the quality and value of all the data in your data resources all at once.
When organizations take an incremental approach to implement formal data disciplines, they must mature through experience and incrementally improve their management and governance techniques. Initial use cases for data governance, data management, and data quality programs, should begin by focusing on selected CDEs … and then address additional CDEs, and then additional … you get the picture. Learn from experience and improve with each occurrence to cover an expanding amount of data that is critical to your organization.
Let’s start out by defining a data element and then tackle what it means to be critical. A data element is defined as:
- Any atomic unit of data defined for processing with precise meaning. In other words, databases, tables, and files (even spreadsheets and reports) have many pieces (units) of data. Each singular piece is considered a data element. For example, a customer address can be considered a singular data element or a collection of data elements – street, city, state, territory, country and so on.
- Defined by size (in characters) and type (alphanumeric, numeric only, true/false, date, …). Every data element has specific characteristics that represent how that piece of data is stored in the data resource. For example, a country code may be a text field with a defined length that pulls its allowable values from a reference table of country codes. By only permitting selected codes from a reference table, the element is assured to be standard and consistent in length and value.
A data element is a singular piece of data. Just one. The most atomic unit of data there is. Examples include: account number, social security number, birthdate, dollar amount, and the list goes on forever.
The truth is that there may be many data elements that are considered critical to your organization. The question always arises as to how many CDEs should be focused on at one time. There is not a perfect answer to this question. Some organizations start very small (3-5 CDEs). Other organizations begin with a dozen (or dozens of) CDEs. Others even still have started down the path of focusing on CDEs numbering as many as fifty or more. Again, consider that there will always be more data elements.
A proper number of CDEs to start is one that is reasonable to demonstrate value to the organization given the availability of time, resources, expectations, and required effort. Typically, it makes sense to group related data elements together when applying data disciplines. Grouped examples includes mailing address elements (i.e., city, state, zip), person name elements (i.e., first name, middle initial, last name), or related elements that make up a specific transaction (like a sale, or an event occurrence).
In many cases, the importance of data depends on who you ask. What is critical to one person, or even an entire department or division, may not be important or used by other people or departments or divisions. Often, people within a single part of the organization access their data from different resources or use different data to complete their functions.
There is no single definition of what data is critical to your organization. But it would be safe to wager, that given a standard criterion that assists you to focus on important data first, that not all your data should be considered critical.
Consider the following rules of thumb for determining if a piece of data is critical, and thus a CDE. Several of the items on this list were shared with me through discussions with clients, while others are general ideas for determining which data elements are CDEs.
A data element may be identified as critical if the singular piece of data is:
- Noted as being critical or protected by organizational policy.
- Considered to be “connective tissue” between information systems.
- Considered to be the “grout between the tiles” or an element that is necessary to improve the meaning and usefulness of other pieces of data.
- Used as a key performance indicator or KPI based on and substantiated by this element.
- Data that is key to the business.
- An element that helps the organization to prioritize its work.
- Associated with regulatory fines/penalties and/or compliance violation risks.
- Associated with significant financial impact risk, such as increased liabilities, costs, or penalties, as well as reduction in assets, revenue opportunities, or profits.
- Associated with interruption or significant reduction of critical business process risk, for an extended period.
The DAMA chapter in The Netherlands (DAMA-NL) documented a process that organizations can use to quantify the criticality of a data element. The process starts by determining the factors (like those listed above) for the selection of CDEs.
Organizations are past the point of recognizing data as an asset. Organizations are, at a minimum, beginning to focus on the true value of data to their ability to thrive or just survive. Organizations are incrementally implementing data strategies or data programs, whether those programs are focused on data governance, data management, or data quality, or some other aspect of the getting value from their data. Since it is impossible to “flip a switch” and all-of-the-sudden have the same level of formal discipline associated with all the data in the organization, it is important to have a method to prioritize the data that will be in focus. I hope that this article will help you to determine which data are critical data elements to your organization.
