December 1, 2007
By James Kobielus,

Mastering Customer Records

Part 1 of a four-part series.

CRM needs services-oriented architecture (SOA) to realize its potential. To grasp that point, you need to understand SOA, but you also need to see how CRM has, from the very start, relied on an information governance discipline that is, essentially, a form of SOA.

SOA focuses on maximizing the reuse, sharing, and interoperability of networked corporate resources. SOA is often positioned as an application integration paradigm, but it also applies in full force to another critical corporate resource: data. For any enterprise, one of the most precious resources is master data -- in other words, the official systems of records, such as customer profiles, product configurations, and financial accounting information, upon which the business runs.

Master data management (MDM) is essentially a body of practices that realizes the aims of SOA in the realm of data management. In a well-designed SOA-based MDM environment, a user knows she can rely on information that is maintained in her company's reference data stores -- no matter how many repositories there are, where they reside, or what applications are used to manage them. This is because all that precious content has been transported, consolidated, cleansed, and secured in keeping with official corporate policies, and by a common set of official corporate data management tools.

CRM was founded on what is basically the same notion: that master customer records should be consolidated into a central database controlled by a single administrative application. However, as we all know, that ideal has rarely played out in reality. In most organizations, master customer data is scattered over diverse applications and databases. Customer profile, account, and transactional records are subjected to a fragmented, inconsistent, rickety set of manual and automated administrative processes. Multiple versions of the same customer data may -- and often do -- permeate many organizations. Customer records are often out-of-date and riddled with errors.

It was exactly this messy situation with real-world CRM that spawned the adoption of MDM's killer application: customer data integration (CDI). Many MDM implementations are purely CDI-focused: Customer data from many sources is consolidated into central databases called data warehouses, data hubs, or data marts. These repositories feed enterprise applications in contact center management, sales and marketing, customer support, business intelligence, corporate performance management, governance risk compliance, and other key areas.

But no full-fledged enterprise CRM environment runs purely on customer records. To fully serve the customer, or target the prospect, you need to leverage other master data entities -- product information, accounting and billing records, supply chain and inventory data, and so on -- and you need to ensure that all these records organize data under a common set of hierarchies, such as organizational business units, operational territories, and fiscal calendars.

MDM solutions -- by vendors such as IBM, Initiate Systems, Oracle, SAP, Siperian, Teradata, and Tibco -- support those multi-entity requirements, providing general-purpose infrastructure plus the hierarchy management tools for all classes of reference data. Increasingly, MDM solution providers are offering extensible, multi-entity solutions that support current and evolving requirements, and that play well in the world of SOA-based integration. Most commercial MDM solutions either bundle data quality (DQ) tools or integrate with third-party DQ premises-based or hosted services. Most MDM vendors provide certified, SOA-based integration with providers of solutions for profiling, verification, de-duplication, correction, and enhancement of customer name/address data, which is critical for robust CDI -- and, hence, CRM.

So, clearly, today's CRM applications depend on SOA to ensure that customer data has been cleansed and is truly ready for business. Increasingly, CRM without SOA is inconceivable--or just far too risky.

Click here to read Part 2 of this series, here to read Part 3, and here for Part 4.

James Kobielus (jkobielus.blogspot.com) is principal analyst for data management at Current Analysis.

Every month, CRM magazine covers the customer relationship management industry and beyond. To subscribe, please visit http://www.destinationcrm.com/subscribe/.