|
Becoming the master of your data domain
Eliminate many of your most arduous data-management tasks and create an enterprise repository of all your company's data.
by Ronald Swift
In today’s market, a lot of buzz surrounds the topic of master data management (MDM). Numerous vendors have jumped
on the MDM bandwagon. The problem is that each of them has a slightly different view of what MDM actually is.
To understand the topic, we must first define the term: Master data is simply reference data that is shared across systems and used to classify and define transactional data. This master data differs from both transactional data and metadata. Transactional data is a record of a business event, such as an order, forecast, inventory or a point-of-sale (POS) operation. Metadata is information about data; that is, an information layer defining the data characteristics such as data type, location, etc.
Master data can define business objects such as customer lists, supplier lists, charts of accounts or bills of materials (BOMs). It can also highlight relationships in the way business works; for example, organizational hierarchy, customer hierarchies or product hierarchies. Master data management,
then, is a foundation for providing a single view of the business for that data—customer, product, employee, supplier
or user-defined—across heterogeneous systems. The larger MDM universe contains subsets such as customer data integration (CDI) or product information management (PIM). Customer data integration refers to customer reference data, while PIM covers product reference data.
If we break down MDM offerings into their simplest form, we see two main components. The first component is an application that provides a heterogeneous interface, work-flows, synchronization functionality, business rules and so on. The second component is the repository or infrastructure that provides the underlying master data storage platform. In some cases, the repository may incorporate part of the management layer and the interface
to the data synchronization pipeline, such as through a message bus or replication engine. It may also include the infrastructure to support event-driven processes with triggers, stored procedures and queues.
Some MDM applications simply define the business rules for the event—the supporting repository is what actually detects and reacts to it. In other cases, the repository functions only as a data storage platform, while the MDM application provides the underlying infrastructure to detect and react to the event, as well as to transport the updated master data to the other systems.
MDM in action
Introducing a new part into a manufacturing process or structuring a one-to-one marketing campaign should be a simple process, right? That may have been true years ago when processes were less complicated and companies dealt with comparatively few automated systems and applications.
Today, such seemingly simple acts can
be surprisingly difficult to implement because of sources of confusion like duplicate records, inaccurate financial reserves or revenue management and so on.
Consider a manufacturing company with five plants. The company procures similar
raw materials for each location, which
would potentially allow it to negotiate
volume pricing and fast-payment discounts from suppliers. Over the years, however,
the IT staffs at the various facilities have
each developed their own data systems,
so each plant uses a different identifier
to refer to the same parts or suppliers.
No common reference number exists to translate the enterprise inventory position.
As a result, supply efforts have become impossible to coordinate—one plant pays
a premium for expedited material from a supplier while another plant scraps the
same material as excess inventory. Plants
bid against one another for materials, or pay for a purchase-price variance even as the company as a whole exceeds the cumulative volume threshold. The company can’t leverage its supply base to full potential because it can’t recognize how much business it’s doing with a given supplier.
Collaborating with partners in the supply chain requires accurate, synchronized product information. Effectively managing large volumes of product information throughout the supply chain can help companies increase revenues and decrease costs by speeding new product introductions, automating procurement, improving invoice and order accuracy, enhancing overall business process efficiencies and making it easier to do business with trading partners. Master data management applications provide companies with the tools to get in sync.
In the marketing world, MDM addresses many challenging aspects of CDI—an essential element of customer relationship management (CRM). Customer data integration involves consolidating and managing customer information from all available sources. The process involves a prodigious amount of data, including contact details, customer valuation data and information gathered through interactions like direct marketing. Such information yields a complete view of the customer—at least, it does if it’s accurate and manageable.
Master data management can synchronize such customer data, eliminating many of the problems that might occur in structuring
a one-to-one marketing campaign, such
as listing Fred Smith of Boston as Frederick Smith in one system, Mr. Smith in a second and Fred Smith in a third. To further complicate matters, another system may list him at an outdated address. Such disparities create a host of problems for marketing, not to mention accounting, which may never receive payment from Fred Smith because the bill is misdelivered. Without functions to coordinate customer reference data, campaigns can become wastes of time and money.
Synchronizing reference data from various systems can help eliminate a host of other potential problems. It can reconcile records from the CRM solution with those in the accounts receivable database, for example, or determine why accounting only shows 800,000 customers while the CRM solution holds one million customer IDs.
Fine-tuning MDM
Many architectures are being promoted
for MDM. Time requirements for the update and synchronization of the master records vary widely. One way to bring master records up to date is through bulk processes. In
bulk processes, changes to reference data
are implemented daily, weekly or monthly
via bulk loads to the master repository, as well as to each application’s local copy of
the data.
Business requirements might dictate near real-time synchronization of reference data within operational and analytical applications. In these cases, leveraging the existing message bus infrastructure allows updates
of the local copy of the reference data to be synchronized with the master data, so that local changes become global. The business also needs to access the reference data directly for administrative purposes. This can be accomplished through MDM.
The truth is, companies are going to need vendors to sustain a combination of all architectures, or a hybrid MDM architecture, any of which Teradata will support. Existing applications of most users leverage reference data. They are unlikely to change their existing applications to access a single repository, so an MDM solution must support the ability to update a local copy of the reference data for each application. As companies build out service-oriented architectures (SOAs), MDM solutions must also provide new applications with access
to a master data service, allowing them to directly access and update reference data.
A well-planned MDM solution yields a better understanding of customers and guards against supply-chain and manufacturing process disruptions. It also enhances productivity and efficiency through more accurate and effective use of operational systems. With MDM, businesses spend less time weeding through seemingly unconnected data and more time building revenue or reducing costs by making use of that data, now synchronized across the company network. Teradata’s active data warehouse
is the underlying infrastructure that lets customers leverage their existing data warehouse to perform MDM—on the same platform at the same time. T
Ronald Swift is vice president of cross-industry solutions marketing for Teradata.
© Teradata Magazine-June 2006
back to top
|
Quick Links
Print
Send to colleague
PDF
E-mail us