Why Master Data Management
Fails To Meet Expectations
Learn how to fix your Master Data Management Systems!
Disparate Master Data
Master data can be the “glue” that binds data together or the wedge that drives data apart. Master data that is compatible binds data together while disparate master data separates data and causes data silos. Master data Management (MDM) is an attempt to bind data together by providing a consistent and clean set of master data records for the enterprise. Master data, sourced from disparate data systems, is consolidated and cleansed in the MDM data system to improve the quality of the master data set. This corrected set of master data may then be distributed to update the source data systems; synchronizing the master data values across these data systems. However, traditional MDM fails to provide the most important master data corrections as we discuss below.
Today’s data systems are best characterized as disparate data silos. Disparate data silos are caused by conflicting master data representations as depicted in Figure 1. In this figure, three database tables are represented, where each table contains different representations of the same address master data records.
Figure 1: Traditional Data Design produces Disparate Representations of Master Data
Each database table resides in a different database (Database1, Database 2, and Database 10). These three database tables differ in their representations by the name of the database tables (AddressTable, Addr, and Address), and by the names of the columns in each table. These three database tables each have a different primary key to uniquely identify the address master data records (AddressID, Addr ID Prty ID, and AddressID). The address master data record values are also inconsistent (123 Main St., 123 Main Street, and Main St.). The address master data is in conflict in representations for the above reasons. Please note that first address data records in each of these three database tables represent the same actual address. In addition, the second data records in each of the three database tables also represent the same actual address. Likewise for the third data records in each of the three database tables.
Common sense would dictate that the matching address data records in each of the three database table should be directly relatable. That is, the first address data records in each of the three database tables should be related as they represent the same actual address in slightly different forms or representations. Unfortunately, the ability to resolve the equality of these disparate address master data records by the computer is not possible without software-based data transformations. For example, the primary key data values for the same address master data records are not equal across databases. With disparate master data like that in Figure 1, it is not possible to directly relate (no data transformations) the same address master data records is indicated by the blue double headed arrow covered with the red “x” shown in Figure 1.
Master Data Management’s impact on Disparate Data
Many organizations have turned to MDM to solve their disparate data problems and to eliminate data silos. Unfortunately, they just barely missed the mark! As you will see below, the MDM corrected data is still disparate.
Traditional MDM methods normally extract the master data records from the source data systems. The extracted master data is then consolidated into a clean and consistent set of master data representations within the MDM system. This consistent set of master data is often referred to as the “golden copy” of the master data. The master data record data values from this golden copy is often returned to the original source data systems in an attempt to synchronize the original master data records.
Figure 2: Conflicted Master Data Form after Master Data Management Methods
In our example, the address data records shown in Figure 1 may be used as three sources of master data for the MDM system. The address master data may be processed to form a golden copy of address master data. Figure 2 shows the updated synchronized address master data records that result from the MDM process. While the improvement in the consistency of data record values is obvious (compared to Figure 1), the improvement falls short of what is needed to make the data sharable across data systems. Unfortunately, the conflicts in the database table name, and the table columns and the primary keys, are not resolved by the traditional MDM process. The resulting address master data is still in conflict such that the computer may not join the appropriate matching address master data records.
This is where traditional MDM fails to meet expectations. The data systems that have been synchronized using traditional MDM methods remain disparate data silos. All that time, effort and capital have not corrected the original problem. The good news is that data compatibility can solve this problem.
Compatible Master Data Management
As was shown in Figure 2, traditional MDM will always fall short of making disparate data sharable across data systems. With traditional MDM, the primary key data values and the data structures may not be changed – so the master data representations remain disparate. In contrast, MDM that uses data compatibility methods, that is compatible MDM, solves the problem completely and easily as shown in Figure 3. In Figure 3, three database tables and their address master data records are now a consistent representation across the databases. All three database tables have the same table names, the same database table column names, and the same primary key columns. Also, please note that the primary key values are consistently applied across databases. The primary key data values indicate which data records represent the same address master data record and which data records do not represent the same addresses. It should be obvious that a computer may easily discern which data records should be joined as shown in Figure 3 by the red arrows. Again, these joins are not possible with traditional MDM methods as shown in Figures 1 and 2.
Figure 3: Compatible MDM fixes your data at its source.
Our patented Data Compatibility Products and services are used to convert disparate data silos into compatible data systems that freely share their data with any other compatible data systems. When compared to the master data representations of Figure 2, the Figure 3 master data representation is very clean and a tremendous improvement. Of course, the address master data records in this compatible structure shown in Figure 3 need to be related to the original or source disparate address master data records from Figure 1. The level of data abstraction provided by the Figure 3 database tables allows for the complete data management required by master data management such as deduping any redundant data records.
Maxxphase Data Compatibility Products may be applied to any data system to support the same form of direct data interoperability as is shown in our Figure 3 example. An unlimited number of data systems may be made compatible simply and as easily as demonstrated in this white paper. The ability to link the same data records across multiple data systems is a very significant database feature which is not supported by disparate data silos.