Within an organization, master data is used to identify and describe things that are significant to the organization.
NOTE 1 In cataloguing applications, master data is used to describe things called "items".
Figure 1 depicts a taxonomy of data, showing where master data fits.
NOTE 2 Figure 1 is not intended to be a complete taxonomy of data; it is only intended to show the context of master data.
Figure 1 — Taxonomy of data (for master data)
Master data is typically referenced in business transactions through an identifier. The identifier is commonly a reference both to the thing itself and to a master data record (MDR) that describes the thing. The MDR is commonly held in a central repository.
EXAMPLE 1 It is common for the central repository of MDRs for an organization to be the organization's enterprise resource planning (ERP) or master data management (MDM) system.
NOTE 3 What is logically a single MDR can be represented by several physical records in a software system.
EXAMPLE 2 In a relational database implementation, a master data record could consist of rows from several different tables.
NOTE 4 A MDR that describes something can be identified via a reference using its identifier. Something can be described by characteristic data, represented by property values. Additionally, something can be described by descriptive strings or definitions.
Identifying references are designed to be used as references to master data held by others.
EXAMPLE 3 A corporate tax identifier, an individual's national insurance number, and a part number assigned by a manufacturer to an item of production are all examples of identifying references.
In order for an identifying reference to be meaningful, it shall be associated with a system of identification.
EXAMPLE 4 The organization that issued the identifier can be specified by the metadata, as is common in tax identifiers, but a part number is meaningless if the manufacturer that issued it is not known.
A description can be computer interpretable characteristic data, which is typically represented as property values, or human readable text. Some properties are differentiating. Because of the ease with which they can be processed, numerical or controlled values are most often used as differentiating.
One of the key aspects of managing master data quality is managing duplication. A consistent approach to managing and eliminating inappropriate duplication is a critical part of master data management.
A characteristic that is considered differentiating by one organization could be considered non-differentiating by another organization.
EXAMPLE 5 A manufacturer would have a different master data record for each of its items of production. When, from a buyer's perspective, several items of production (produced by the same manufacturer or different manufacturers) share the same characteristics of fit form and function, the buyer may group under a single item of supply and assign a "stock number" as the identifying reference for the item of supply. In grouping several items of production as a single item of supply, the buyer is making a decision to consider as non-differentiating one or more characteristics that the manufacturer(s) consider differentiating.
A characteristic that is considered differentiating by one function within an organization may be considered non-differentiating by another function within the same organization.
Master data is not necessarily static. Also, the number of characteristics needed to describe something will vary by business function. As the number of differentiating characteristics vary, MDRs may have to be differentiated when characteristics are added or changed to differentiating. MDRs may become duplicates when characteristics are removed or changed to be non-differentiating.
Examples of master data include:
vendor master: This typically describes a vendor in term of its location and legal status. Much of the mandatory data in a vendor master is prescribed by law as it is a common requirement for a company to be able to identify all entities to which it has transferred funds.
customer master: This typically describes a customer in terms of a trading entity. At a minimum it will include the contact information necessary to transmit invoices and may contain confidential information such as credit card information.
NOTE 5 If personal data is maintained in a customer master, it can be subject to data protection legislation such as the United Kingdom Data Protection Act of 1998.
item or material master: These masters typically describe tangible items that are tracked, inventoried or regularly purchased. While they are often restricted to items purchased under contract such as production materials they can also be used to improve the quality of spend analysis associated with maintenance, repair and operations (MRO) purchases. Material masters are also commonly used to support bills of materials (BOM) or to in design where they maybe referred to as common parts catalogue or a preferred part list. A variation of the material master is an illustrated parts catalogue (IPC) or a spare parts list.
item of supply concept: These masters include a reference to an item or material master, plus packaging and quantity information;
service, procedure or process master: These masters are still relatively rare except in the health care and vehicle repair industries where automated billing for services or insurance reimbursement is common. Typically a service is best described as a procedure or a process.
EXAMPLE 6 The American Medical Association's Current Procedural Terminology-4 (CPT-4) codes is an example of a procedure master.
asset master: These masters are commonly used to track items whose purchase price is over a preset monetary value, or whose cost is depreciated over several years. Assets are commonly associated with a unique identifier (serial number) and often associated with movable items where date (time occasionally) and location need to be verified and reported. Correct modelling of an asset master is important to be able to track not only the location and value of the asset over time but also the maintenance and repair activity. A typical problem with asset management is changing specifications over the asset’s life span. Deciding at what stage an asset has been so modified as to require the creation of a newly described asset is often a challenging issue.
location master: Other than delivery services it is rare to see a separate location master, yet separating out the location master from customer and vendor masters typically leads to improved data quality. The data model for a location master is basically simple as in theory it describes a physical location where global positioning coordinates provide the absolute reference. In practice there may need to include other delivery instructions such as a postal address.
point of contact, employee or human resource (HR) master: These masters typically describe an individual. Commonly they include information related to the relationship with the employing organization but these are better treated as transaction data as opposed to master data. They often contain confidential information.
NOTE 6 The data contained in these masters can be regulated and subject to data protection legislation such as the United Kingdom Data Protection Act of 1998.