Services are the building blocks of SOA, and like building blocks of a house or a building, the quality will define the value of the finished product. In this case, the SOA itself. Thus, spending...

Service-Oriented Architecture (SOA) represents a more flexible approach to building IT solutions. Combining SOA with other complementary technologies and approaches further enhances the ability to build more agile solutions. These solutions are easier to change, provide improved access to time-critical information and are more aware of - and responsive to - business events throughout the value chain. Some of these complementary technologies and approaches include: Business Process Management, Business Rules, B2B Services, Master Data Management, Event Processing and Mobile Computing.  The focus of this article is on Master Data Management and SOA.

SOA is an approach to distributed computing that considers software functionality as services on a network. SOA represents the next major step in the evolution of IT strategies. Businesses can look to SOA as the best way to leverage information technology assets and to provide the business the agility required to compete in today’s economy.  Fully leveraging the value of SOA requires establishing a discipline to rationalize the silos of fragmented and stove-piped master data.  Such an approach is known as Master Data Management.

Master data can be defined as the information required to create and maintain an enterprise-wide "system of record" for your core business entities, in order to capture business transactions and measure results for these entities.  Master Data Management is an investment and a discipline that will provide real business benefits. These benefits include improving the quality of data, reducing data management costs, increasing accuracy in analysis and reporting, and leveraging existing investments in CRM, ERP and SRM.

MDM creates a Unified Business Vocabulary for SOA. Complex data models can be reduced from potentially hundreds of tables into a manageable number of components. The components define a vocabulary in terms that are understood by business users.  Representative terms in the Retail industry include Order, Product, Inventory, Location and Retail Sale.  In Insurance the terms include Policy, Policy Holder, Agent, Claim and Location.

The root problems behind poor MDM are fragmentation and duplication.  Master data, such as customer, product and supplier records, are located in multiple, disparate IT systems (CRM, ERP, legacy, etc.) that run independently of each other. As a result, organizations have dispersed and duplicated data that creates multiple views of the same information and inaccurate representations of customers, business transactions, product usages and license agreements. For example, duplicate customer and product records may cause improper revenue recognition because multiple invoices are created for the same customer or product. Bad customer data also inhibits a company’s ability to conduct accurate analyses, leading to flawed customer segmentation and poorly targeted marketing campaigns. An inaccurate understanding of customer interactions limits sales effectiveness and results in wasted marketing dollars. For many companies, regulatory legislation like Sarbanes-Oxley compounds the problem, because unclear revenue and profit reconciliation increases reporting risks. Unless a company invests in reconciling its MDM, poor data will continue to drive down customer service levels, perpetuate an inaccurate view of business operations, and increase an organization’s exposure to compliance issues.

The disjointed management of master data results in inaccurate views of a company’s business, leading to millions of dollars in lost sales and cost savings opportunities annually.  During the last 2 years, retailers’ mandates for data synchronization have shed light on the immense problem of poor product information across the CPG and retail supply chain. Siloed operations and disparate systems perpetuate the problem of maintaining a centralized view across the enterprise. As companies try to publish and subscribe their product information to share with partners, they are finding that achieving a consistent and accurate view of product information inside the organization is a more strategic undertaking than originally thought. CPG and retail companies are not alone; duplicate and inconsistent data on products, customers and suppliers is a significant issue that plagues a number of industries, including high tech and financial services.

The MDM solution unifies the data silos into a Unified Data Architecture.  It is a modular system, with simplified connectivity and enhanced scalability for both intra- and inter-enterprise functionality.  It is easy to implement with little impact on existing operational systems.   It is business-user accessible and business-user driven for validation.  It also supports complex data models. It consolidates master data from disparate IT systems, interfaces with enterprise data warehouses, maps between multiple data views and preserves master data history.  A unified data management provides the following benefits:

• Improves the quality of data
• Enables more centralized control over information
• Ensures cross-system data consistency
• Enables dependable cross-system, enterprise-wide business processes and analytics
• Reduces data redundancy and data management costs
• Allows the sharing of a central repository of information from applications
• Allows accurate analysis, reporting and sourcing

MDM as a discipline is becoming important because of the shift to service-oriented architecture (SOA). By 2009, more than half of today’s leading companies will redesign their enterprise applications architecture to base it on SOA. This best practice architecture approach requires MDM as a first step. The rationalization of services yields a single shared service that provides a system of record for data of a certain type (such as customer or product). This requires a corresponding rationalization of the underlying data. An MDM solution separates data from business processes and applications enabling the data to be rationalized. MDM creates a flexible system that can support the migration to SOA. The move toward decoupled software services means that organizations will require a centralized data or metadata hub which MDM provides.

Tom Dwyer is the VP of Research for BrainStorm Group, the Editorial Director of BPMInstitute.org, and Editorial Board Member for SOAInstitute.org and a co-chair of BrainStorm Group's SOA conference series. He writes, presents and consults on topics that include Service-Oriented Architecture, Business-to-Business, Enterprise Application Integration, and Business Process Management. Mr. Dwyer has conducted primary research and published extensive reports on the Application Software Infrastructure markets.

About Tom's SOA training course.

Back to Articles, including SOA, BPM & BA