Metadata are the data fields where data are created, reviewed, updated, and deleted. We will use a customer profile as an example. Figure 8.1 listed the various metadata fields used to describe a customer profile. These included account name; company name; first name, last name; e-mail address; phone number; billing address; shipping address, city, state, zip code country, province, postal code; the Dun 8c Bradstreet Universal Numbering System (DUNS) DUNS Legal ID, DUNS Site ID, payment terms, sales representative; territory ID, credit limit. Information from various customers within these fields is in the same format but varies because different customers will have differ universal customer identification (UCID) numbers and information.

Customer account metadata is organized as a hierarchy that enables an organization to structure data relationships for process control and reporting. Figure 10.2 is an example for accounts and uses DUNS metadata. The hierarchy of a customer account is important for reporting and understanding customer account relationships such as invoicing, services, shipments, and other functions. In a large, multinational organization with several international subsidiaries, reporting will be aggregated to the legal UCID, and lower levels of an organization will have unique numbers. These include a global ultimate UCID, a domestic ultimate UCID, a subsidiary ultimate UCID, and a site UCID for locations within a subsidiary. An example would be different manufacturing facilities in a corporate division within a specific country. This metadata field is useful if purchasing or leasing equipment with warranties or other entitlements associated with a specific machine. This ensures service technicians and remote support teams provide services and parts to the correct machine at a location. If an organization deviates from this hierarchy (e.g., because of poor database design), it will have difficulty merging and sharing metadata with other organizations because the hierarchal relationships will be different. From a process engineering perspective, the consequences of poor hierarchal design will be costly data collection and poor reporting and process control across the IT ecosystem.

Business data stewards measure the quality dimensions of metadata to ensure accurate billing, deliver)', invoicing, and other workstreams. Table 10.3 describes these dimensions and common data quality rules. Most people


Customer data taxonomy. DUNS = Dun & Bradstreet Universal Numbering System.

TABLE 10.3

Data Business Rules

Metadata Field

Business Rules

Company Name

UCID legal name corresponds to a physical address. Name not abbreviated.

Name must not be blank.

Name cannot be all numbers or special characters.


UCID site must be a physical address or post office box. Address must not be blank.

Address cannot be all numbers or special characters.


E-mail has an @ symbol.

No blank spaces.

E-mail has a valid domain extension such as .com, .net, etc.

are familiar with a few of them (e.g., accuracy, completeness, etc.), but from the perspectives of data governance and process efficiency, all of them are important. To better measure these dimensions, data quality rules are applied to enable automatic verification of each dimension to its rule. As an example, the UCID legal name must exist, it cannot be abbreviated or be blank, and it cannot be all numbers or special characters. The other data quality dimensions would also be applied to this metadata field. Data quality is reported using a format shown in Table 10.4. Reporting is stratified to understand data quality for location, account, source system, and other relevant factors.

A business process from a data perspective is an aggregation of metadata from several sources that build work products such as a customer order profile. How an organization’s systems use metadata is important for information governance. Data mapping or data lineage traces the sources of metadata and how it flows through various IT applications to create work products, such as an order or invoice. Data lineage shows whether metadata came from a trusted source. It is also important if there are quality issues or if the business rules need to be modified. Software tools are used to crawl though IT applications and trace the end-to-end flow of metadata. Visualization of metadata lineage is important to see high-volume transaction flows to focus process improvement efforts, including the evaluation of metadata performance against business rules. In summary, because of the number of applications and platforms used by organizations, process improvement professionals need to understand metadata quality, lineage, and governance to work effectively with business data stewards, business stakeholders, and others to improve process quality and efficiency.

< Prev   CONTENTS   Source   Next >