In my article in the September issue of DM Review, I talked about data management (DM) in a service-oriented architecture (SOA). That article focused on the DM framework that must be built to implement an SOA. I've received quite a lot of positive feedback, but also a few suggestions about refocusing the direction of the DM discussion on master data management (MDM) - in particular, focusing more on each part of the DM framework: process, technology, architecture and standards, organization and governance. In response to those suggestions, I've decided to write a series of columns on MDM, in the context of the larger DM framework I've already developed. The first part of this series will focus in-depth on MDM processes.
Every company has a core set of data elements used by many or most of the processes and information systems in the company. This is master data. It provides information on the company's products, materials, vendors, customers and finances. Consider data about materials used to assemble products, for instance. For materials, master data elements would be "part name," "part number," and so on. This data spans business processes such as program management, product engineering, order management, materials procurement and product manufacturing - just to mention a few. This type of enterprise-wide core data is critical to the company's success, and it must be managed effectively.
The first step in effectively managing master data is developing optimal business processes at the enterprise level. A process model is needed that reflects leading practices for key processes in a given industry segment. The process model should be depicted in three or four levels of detail with decomposition and flow diagrams. The model also should be mapped to enterprise software packages to assist in technology-enabled business transformation. Finally, the model should be delivered in common graphics and database tools such as Microsoft Visio and Microsoft Access. If you don't have such a process model in place, it is critical that you develop and implement one before moving forward.
Once your business processes are in place, you can begin to develop processes to manage data that is used by those processes. There are four processes essential to a top-notch MDM strategy:
- Data migration and integration
- Data maintenance
- Data quality assurance and control
- Data archiving
Let's begin with the data migration and integration process. This is where data from legacy or external systems is aggregated, cleansed, mapped, converted, loaded and integrated into your information architecture. For this process to run successfully, it should have a detailed data cleansing routine predefined and in place. The process should weed out duplicate and obsolete data, as well as correct any errors in the data. The mapping task within this process should identify and remedy any gaps in the incoming data. Finally, the conversion and loading process should perform - according to predefined functional and technical specifications - a thorough translation and verification routine before the data is loaded into the system of choice.
Next is the data maintenance process. Data maintenance is simply the ongoing synchronization of your master data definitions with your company's business processes, roles, service level agreements (SLAs) and automated processes. This is where you confirm that any adds, changes, deletes and/or archivals of business processes are synchronized with the master data definitions. For instance, if you add a four-digit extension onto the ZIP code in your information systems, your master data definitions should reflect that. This is also where any changes to SLAs (such as performance requirements, change request information, etc.) are implemented at the master data level.
Next, let's look at the data quality process. "Data quality" is a catchall term used for the level of correctness, consistency, completeness and integrity of your company's data. Simply put, this process focuses on the control and reduction of error while data is moving through your information systems. A good data quality process will have strict control measures in place not only to evaluate and correct, but also to continuously improve the quality of the company's data.
The last process is data archiving. There is really not much to say about this process except that Uncle Sam has caused it to be very important. With the advent of new, more extensive reporting requirements, it is incredibly important to have a process in place that synchronizes and maintains information on legal, tax, audit and operations matters for the time periods required by various governmental agencies such as the SEC and IRS. Here, clearly defined functional and technical specifications are needed that make the stored data easy to find and retrieve when needed.
I admit, data management is not the most scintillating topic in the world, and I'm not the first (or last) to write about it. However, as sleep-inducing as the peculiarities of data migration, quality, maintenance and archival might be, they are critically important to providing accurate, consistent data to decision-makers. Data accuracy and consistency spell the difference between success and failure. That alone should be enough to keep you awake.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access