As Published In September/October 2006

Oracle Data Hubs centralize and synchronize the truth of your enterprise.

It's a classic courtroom drama. Actor Jack Nicholson, playing the besieged Col. Jessep in the movie A Few Good Men, is directed by Tom Cruise's Lt. Kaffee to "tell the truth." From the witness stand, Jessep glowers at Kaffee and snaps, "You can't handle the truth!"

Obviously, Jessep wasn't talking to a business executive. Businesspeople not only can handle the truth, it's something they thrive on. It's the incomplete, duplicate, fuzzy, and slippery facts that pass as corporate truth that drive executives crazy.

For years, companies of all sizes and industries have been challenged to find ways to create a "single source of truth" about their customers and products, often with little success. That challenge has now been made easier because of Oracle's expanding family of data hub technologies. Data hubs provide a central place where companies can store customer and product information that's been standardized, enriched, deduplicated, and made available to any enterprise resource planning (ERP), customer relationship management, engineering, or other production system that needs the information. Gone are the frustrations of inaccurate records and not knowing if Jennifer Jones and J. Jones are different clients or one and the same.

For example, IHOP, the iconic, nearly 50-year-old family restaurant chain, began a comprehensive effort in 2003 to replace manual business processes and knock down departmental data silos. "Each department had its own view of who our customers were," says Jeff Valine, director of IT applications and project management for the Glendale, California, company. These views ranged from restaurant guests to the nearly 400 franchisees that operate more than 1,200 IHOP restaurants in North America. "If you asked three departments for a list of customers, you would literally get three different lists."

Businesses find that data discrepancies do more than spawn frustration, they also hurt customer service. "If companies have information stored in multiple places, they can't coordinate activities intelligently," says Bill Swanton, vice president of research for AMR Research, in Boston, Massachusetts. "The last thing you want is to have a salesman pitch a new product to a customer who has just finished making an irate service call."

Reducing Errors, Duplications

For IHOP, just as bad as the departmental data discrepancies were the inefficiencies and errors that cropped up when the company's franchise development and legal departments stepped candidates through the franchise qualification process. Business development people maintained franchisee information in a desktop database before transferring relevant details in paper documents sent to legal, which promptly keyed the information into a spreadsheet. "There was a lot of duplication and chance for input errors," Valine says.

Then IHOP made Oracle Customer Data Hub, part of the Oracle Fusion Middleware product family, a core component of its technical reorganization effort. Now, with a customer data hub in place, "the data flows from one department to the other without any paper being generated," he adds.

Earlier this year, IHOP launched a custom application on top of the data hub that allows the company to track milestones for new restaurants, including major steps such as real estate approvals and construction timelines. IHOP built the program in Java using Oracle Application Server. "Once you've got that data hub infrastructure in place you can quickly tap into that data and bring up a new application," says Valine.

Snapshots IHOP www.ihop.com Location: Glendale, California Industry: Hospitality Employees: 897 Oracle products: Oracle Database 10g, Oracle Database 9i, Oracle E-Business Suite 11.59, Oracle Customer Data Hub, Oracle Customer Relationship Management Suite, Oracle TeleSales, Oracle Property Manager, Oracle Sales Online, Oracle Collaboration Suite, Oracle Portal Glacier Garlock Bearings www.garlockbearings.com Location: Heilbronn, Germany Industry: Industrial manufacturing Employees: 1,200 Oracle products: Oracle Product Information Management Data Hub, Oracle Order Management, Oracle Inventory, Oracle Purchasing, Oracle Manufacturing, Oracle Financials

For Valine, the success of IHOP's data hub isn't only measured in return on investment dollars, but by knowing the company is built on the right business foundation. "It's really about improving the efficiency of the overall corporation," he says. "We believe this is the right infrastructure to have in place."

Implementing a data hub typically follows clearly defined steps. First, companies choose the core identifying attributes that ensure a unique occurrence of a master record. This is usually a collaborative effort among departmental users, privacy officers, and IT. For example, attributes in a business customer's identifying data often include the company name, corporate Web site address, master phone number, and physical address.

The second step is the easiest. What information (data elements) does the company want to store about each master record? In larger application portfolios, each source system has some elements of truth to contribute to the master record. For example, the accounting application may contain the single source of truth for the bill-to address, but the service application contains the single source of truth for the installed-at address. The Data Hub technology lets companies source each data element from each data source if necessary to build the composite master record. The single-source-of-truth record is further supported by a metadata framework that allows companies to personalize the data models by adding unique attributes to them without programming.

After the record is uniquely identified, a robust rules-based matching engine determines whether the record is a duplicate. If it is, then the hub doesn't create a new record in the data registry. Instead, it adds a record in a cross-reference table.

Data models structure and store the data, but the next step is to ensure the information itself is accurate, up-to-date, and complete. "For companies that are new to managing data quality, often the best strategy is to take an important subset of the total records and develop the data quality business processes and corresponding matching, validation, and enrichment rules," says Peter Heller, senior director of applications product marketing at Oracle. "Most companies are surprised at how bad their data quality is, and very early into the project they realize the need for dedicated 'data quality' resources. Data hub best practices call for a central data librarian supported by a data governance committee from various lines of business to centralize data quality business practices, hub operations, and manage the complete data lifecycle," he says.

Doing More

Large companies—and even smaller companies—should do more to manage data. "Even if a company runs a single enterprise software package, each department that touches customer information needs to coordinate its data management processes—and most companies don't," says Heller. "And the greater reality is that most large organizations run dozens of business systems with customer and product information, and the ability to successfully cross-reference that data has become an essential competitive advantage." Enter data hubs.

Hubs not only store and maintain accurate records, they provide an integration layer for synchronizing the data among information sources. Hubs based on open standards, such as Oracle's, work with any combination of applications from various vendors. Further, workflow capabilities in Oracle's data hubs automatically route information from department to department for sign-offs, which can reduce approval times from days or weeks to hours, AMR Research's Swanton adds.

Oracle has unified its customer data management across Oracle's portfolio of applications and product lines: Oracle Customer Data Hub has been integrated with Oracle's PeopleSoft Integrated Customer Master and soon will be integrated with Oracle's Siebel Universal Customer Master. "Soon Oracle customers running mixed environments will be able to create a customer record in one product line and have it automatically synchronized with each of the others," Heller says.

The Oracle Product Information Management Data Hub (Oracle PIM Data Hub) addresses some of the most nagging problems within engineering and manufacturing companies—accurately managing parts data. Some part designs are in constant evolution, while other standard designs remain constant but are prone to duplication when inaccurate records make it difficult for engineers to determine that the parts already exist.

Maintaining associated engineering drawings and materials hazards sheets for parts is another management problem. "The bulk of data around products comes from a wide variety of sources. Some of the information is highly confidential, some of it globalized, and most of it requires strict audit and controls. Worst of all it is largely free-form, unstructured information, making it hard to coordinate, version, and communicate consistently," Heller says. "Consequently it's an important feature that the Oracle Product Information Management Data Hub can securely handle this diversity of information."

Data reliability for parts and materials is important for Glacier Garlock Bearings (GGB), a division of EnPro Industries and the world's largest manufacturer of metal-polymer plain bearings. GGB manages 15 sites worldwide from its Heilbronn, Germany, headquarters. As the company evolves its lean manufacturing strategy, it relies on universal part numbers to eliminate the duplications and inconsistencies that hampered parts management in the past. "We looked around and saw that the [Oracle] PIM Data Hub could help us solve this issue," says Matthias Kenngott, IT director at GGB.

GGB now uses Oracle PIM Data Hub to manage all of its parts and synchronize the data between the Oracle hub and three ERP systems. "Oracle is now the master for all product data information," Kenngott says. "It stores data in a format that the other ERP systems understand."

By culling duplicate part numbers, Oracle PIM Data Hub helped GGB reduce the master list from 50,000 to about 30,000 items. Even so, only about 2,000 standard parts that could be easily identified by their attributes existed among the entire 30,000-item collection. GGB had to rely on engineering drawings to identify the rest. However, the Oracle PIM Data Hub keeps the relationships between parts and drawings manageable, even when parts are modified. The hub is "one single place where we store our information about a product, but also attach our drawings for it. We can find out everything about a part from anywhere in the enterprise," Kenngott says.

Next Steps READ about data hubs LEARN more about data management "Data Management Dynamics: the ROI from Data Quality" "Common Sense Guide to Customer Data Management" DISCOVER Oracle Fusion Middleware

The hub also makes sense of the frequent changes made to parts. "When you modify an existing part, you create a link [in the Oracle PIM Data Hub] to the subsystem with the old part number and cross-reference the new part," says Kenngott. The PIM Data Hub simultaneously updates the part reference in the old system. For new parts, GGB just creates a new part number in the Oracle PIM Data Hub. "If it is created in Oracle, it is automatically available in old systems as well."

Now, if a GGB facility doesn't stock a requested part, it can search the hub for parts stocked at any other GGB site throughout the world. In the past, a GGB facility may have ordered the creation of the requested part, assuming one with those specifications didn't exist, which was a prime cause of part duplications. The Oracle PIM Data Hub "definitely saves time and eliminates duplication of effort and redundant entries for the same part," Kenngott says.

GGB plans to connect its suppliers and automotive customers with the new parts management system. And if the company continues to acquire other firms, something it has done frequently in the past, the new Oracle PIM Data Hub-based infrastructure will be a plus.

"If the new company has additional ERP systems, we only need to define the data that will flow between the Oracle PIM Data Hub and their new applications to involve them in our strategy and tools," says Kenngott. "We're well positioned for the future."