|
Data powers the information economy, and one way to make sure information systems are tuned for optimum performance is to utilize data hubs. A data hub is a unified, reconciled, single source of high quality data. It establishes a master identity for a given business subject, such as product, location or customer, and it utilizes a centralized data model to unify common data elements across numerous information systems. Most data hubs, created via a hub-and-spoke model, enable various business applications to "subscribe" to the data hub as a service.
Learn how GGB, the world's biggest manufacturer of metal-polymer plain bearings, used Oracle technology to create a product information management (PIM) data hub. The hub provides a consistent, single source of product data so that the firm can cohesively manage product information in a single location.
GGB utilizes Oracle Product Information Management Data Hub to group products by attributes, and GGB likes the way the application automatically synchronizes attributes. Oracle unifies fragmented data without disrupting current business processes. The benefit is that Oracle Product Information Management Data Hub can centralize, cleanse, and enrich product data. It can also continuously synchronize it with all related source systems. The data hub manages information about 50,000 different machine parts and synchronizes the data among various ERP systems.
|
|
|
Executive Guide to Technology
Data Powers the Information Economy
By David Baum
Maintaining high-quality data is the key to getting the most out of your information systems.
If data powers the information economy, then data hubs keep those information systems tuned for accurate performance.
Consider customer data. Complete and accurate knowledge about customers is what propels most marketing strategies, as managers probe the nuances of customer satisfaction, loyalty, and profitability. Yet executives complain that they don't have a firm grasp on precisely who their customers are, what they are buying, and what else the company might sell them. In recent years, questions of accountability have surfaced as well: which customers put a company's financial reporting at risk? Sarbanes-Oxley mandates require companies to disclose the material risk to revenue and profit, both of which are customer-centric.
Product data is another prime example. Accurate, integrated product information has become a key strategic asset that contributes to the success of many businesses. Yet companies in just about every industry face pressing issues with respect to consolidating, storing, cleansing, synchronizing, and exchanging this vital information. Individual departments have built or purchased information systems to manage product information and automate the associated business processes, such as procurement, assembly, and sourcing. These information systems segregate product data into information "silos," none of which offers a complete view.
In some respects, the growing sophistication of our information systems has become our undoing. Many companies have deployed enterprise applications to manage everything from taking orders to managing finances to logging service requests. Each of these applications tracks data about customers, products, and other crucial information. Each may have perfect data quality. But none maintains an enterprise view.
According to Peter Heller, senior director of worldwide marketing at Oracle, the software industry has offered many solutions to this problem, such as consolidating information systems, deploying data-integration technology, and adding a layer of analytics on top of business applications. "Each of these approaches solves a portion of the problem," he says. "But without an enterprise perspective, that which is gained is quickly lost."
Heller believes the solution is to map out a single view of the domain at hand that can inform each and every business process. To do that, managers need a unified, reconciled, single source of high-quality customer data. They need a data hub.
Tuning in to Quality
A data hub is a data-quality solution that establishes a master identity for a given business subject, such as customer, product, or location. It uses a centralized data model to unify common data elements across multiple information systems. Most data hubs are implemented using a hub-and-spoke model, allowing various business applications to "subscribe" to the data hub as a service.
Constructing an enterprise data store of this type is what motivated GGB, the world's largest manufacturer of metal-polymer plain bearings, to deploy a product information management (PIM) data hub. GGB uses Oracle technology to create a single, consistent source of product data—an enterprise product repository. This type of centralized data hub allows the company to manage product information in a single location in a cohesive way.
"We have created five catalog categories in Oracle E-Business Suite for each subsystem and entity, where we can store special attributes for each site," says Matthias Kenngott, IT director at GGB. "Oracle Product Information Management Data Hub automatically synchronizes the attributes for us."
GGB uses Oracle Product Information Management Data Hub to group products by their attributes, including metal-polymer bearings, thermoplastic materials, fiber-reinforced plastic composite materials, monometallic materials, and special materials. Oracle unifies fragmented information without disrupting existing business processes—centralizing, cleansing, and enriching product data and continuously synchronizing it with all related source systems.
Driving Industry
In many cases, PIM initiatives are driven by operational imperatives that are unique to a particular industry. In retail, for example, as RFID becomes increasingly prevalent, companies are being deluged with new product data, revealing costly flaws in their product data management practices. High-tech manufacturers and distributors face their own set of product management challenges, particularly as new waste reduction and disposal regulations force them to gather, organize, and communicate new types of core product information related to toxic materials.
The best way to manage these industry-specific challenges is to create a single, consistent source of product data—a master repository. This type of centralized data hub allows companies to manage product information in a single location in a cohesive way. By delivering a 360-degree view of products, companies can improve customer service levels, reduce costs, and improve overall profitability. Without centralized data, they must maintain data in multiple locations, increasing the likelihood of poor data quality and operational disconnects that can lead to costly inefficiencies.
Snapshot
GGB
www.garlockbearings.com
Location: Heilbronn, Germany
Employees: 1,200
Oracle products and services: Oracle Product Information Management Data Hub; Oracle E-Business Suite, including Order Management, Inventory, Purchasing, Manufacturing, and Financials
|
Getting Lean
At GGB, the catalyst for adopting this type of hub-based infrastructure is a worldwide shift to a Lean manufacturing model. As Kenngott and other GGB IT professionals drafted a strategy for the associated process improvements, they knew they would need to make some major changes to their company's core information systems. "We wanted to reduce the number of ERP [enterprise resource planning] systems, and we wanted each part to have its own unique part number, no matter where it is handled in the enterprise," he says.
GGB now uses Oracle Product Information Management Data Hub to manage 50,000 parts and to synchronize the data among multiple ERP systems. "Oracle is now the master for all product information," Kenngott adds. "It stores data in a format that all the ERP systems understand."
Setting up GGB's data hub was straightforward, but configuring it to reflect the company's considerable product line was a significant task. "The most time-consuming part of the Oracle implementation was bringing all our sites together and creating a standard for how part numbers should be defined," Kenngott says. "We reduced our ERP systems from 15 to 3, and we reduced 50,000 parts down to 30,000 by identifying overlap."
Creating and modifying all those part numbers manually would have been an immense task, but GGB used the Oracle system to automatically generate part numbers. "We have some complex parts with 20-digit part numbers," Kenngott continues. "Those can be very difficult to create manually. Now, our Oracle Product Information Management Data Hub solution prompts users to enter attributes, and then it generates the correct part numbers. It's faster, easier, and more accurate."
Boosting Efficiency
Today, GGB is hard at work centralizing its massive parts list. Kenngott is confident that the effort will be worthwhile. "The new system definitely saves time," he says. "For example, it eliminates duplicate effort and redundant entries for the same part."
Previously, if a customer asked for a part that a local plant didn't have, operators would create a new part number from scratch, even if one of GGB's sister companies had already defined the item. Today these redundant efforts have been eliminated. All GGB companies can find the same part, with the same price, anywhere in the enterprise.
Enforcing consistent business practices in this fashion makes the entire operation more efficient. "We have reduced the number of times a part needs to be touched, and we are enabling additional process improvements as we move toward a Lean manufacturing model," Kenngott concludes. "We also have an open interface that will make things much smoother if we acquire another company in the future. If the new company has additional ERP systems, we only need to define the data that will flow between the Oracle Product Information Management Data Hub and the new applications to involve them in our strategy and tools."
Trendwatch Q&A
Defining Terms: A Primer on Data Quality
Profit spoke with Ronda Krier, director
of data management for Oracle Applications strategy, to learn the fundamentals of data quality and how today's data hubs can help you achieve it.
PROFIT: Why is data quality important?
KRIER: A fair number of enterprise applications fail because of poor data quality—not because the applications lack the necessary capabilities or the implementers didn't do a good job, but because the underlying data is inaccurate or compromised. Bad data yields inaccurate results, and before long users don't trust the system. They are much more likely to use an application if they know it delivers information that is accurate and complete.
PROFIT: How do companies end up with so much bad data?
KRIER: The root cause is duplicate, inconsistent data generated by independent applications. When there is duplicate data—or data that is classified inconsistently or assigned incorrectly in a reporting hierarchy—transaction activity becomes hidden in ways that are too complex to track and understand. This impacts everything from financial reporting to corporate governance. Duplicate data must be eliminated at the source.
PROFIT: How do data hubs do this?
KRIER: Data hubs rely on enterprise-level data-management procedures to centralize, deduplicate, and enrich data. They continuously synchronize a master data store with all source systems. For example, a customer data hub would maintain complete records of all customer activity.
PROFIT: How does a data hub help workers on the front lines?
KRIER: Let's continue with our customer-data example. With a hub managing the master customer identity, every department can capture transaction activity against the same record. This improves reporting accuracy, employee productivity, reliability of customer analytics, and day-to-day customer relationships. If an accounts-receivable clerk and a salesperson both access the same customer record, they will always have consistent information. If one of them updates a mailing address, both of them know it immediately, because both of their applications are looking at the same information. This can feed lots of other systems too. If I have a consolidated customer master, for example, I can easily pull up pending orders or service requests.
PROFIT: How did data hubs evolve?
KRIER: Back in the 1990s, companies implemented best-of-breed software solutions because there really weren't any comprehensive application suites available. They also acquired companies that had their own types of applications. The result was an architecture that was very heterogeneous. Each application maintained its own information, which meant customer data, product data, and other critical information were scattered across multiple systems. A data hub is a way to create a master data store that multiple applications can draw from, and to view transactions from all these systems in a consistent way.
PROFIT: What is Oracle's answer to these data-quality issues?
KRIER: Oracle has always advocated having a single data model because we believe the first step on the journey to customer insight is to build a solid foundation. For example, Oracle applications handling financials and supply chain interactions all rely on a master customer reference. The Oracle Customer Data Hub and Oracle Product Information Management Data Hub provide a unified, reconciled, single source of high-quality customer and product data, respectively.
PROFIT: Are there other types as well?
KRIER: Soon after the release of Oracle Customer Data Hub, customers started asking us to develop similar functionality for other domains, such as product information and human resources information. We work with one customer that is, because of its unique license structure, creating not only a customer data hub but a location data hub as well. They not only need to see locations in the context of customers but locations as standalone entities, since one property might contain multiple licensing opportunities. If a new resident moves in, the company might be able to extend his or her warranty contract because they have history of usage in that location. This same reasoning applies to other industries and domains—from appliances to windows to air conditioning units.
PROFIT: What kind of return on investment can customers expect?
KRIER: Data quality isn't cheap. Some companies outsource data-cleansing
activities to vendors that standardize data, eliminate duplicates, and supplement it with credit scores and other information. Given the magnitude of these efforts, it makes sense to bring data together in a master environment and then share it with all of your other systems, instead of having to manage it in each individual system. The Oracle Customer Data Hub lets you manage the customer data lifecycle centrally through a workbench that handles all data-quality issues, such as identifying and resolving duplicates, answering user requests for deduplication, updating customer records, and enriching those records with public data from Dun & Bradstreet and other services.
Bad data can also result in lost revenue. For example, when a salesperson calls on a customer, does he have the latest information on the products the customer has purchased? Does he know which service requests have been logged recently? This knowledge, which comes from a current customer master, is the foundation for providing good customer service and growing revenue.
|
David Baum is a freelance writer and a frequent contributor to Profit: The Business of Technology.
|
