Master data management evolves from software technology to enterprise strategy

Like the people who use its products, premium golf club maker PING constantly works to improve its performance. In recent years, the company slashed the time it takes to get new products to market from two years to nine months, while also shrinking delivery time for a custom-made club from six weeks to two days.

By William Atkinson, contributing editor March 1, 2007

Like the people who use its products, premium golf club maker PING constantly works to improve its performance. In recent years, the company slashed the time it takes to get new products to market from two years to nine months, while also shrinking delivery time for a custom-made club from six weeks to two days.

By next year, PING expects customers to be swinging its custom-made clubs 24 hours after placing an order.

This Phoenix-based manufacturer is able to continuously set—and reach—new performance goals largely because it has mastered the art of managing its corporate data. “We have a single database that we use both for day-to-day operations and decision support,” says Kent Crossland, PING’s director of information systems. “[That database also supports] everything that has do with our interaction with a customer.”

The database Crossland refers to actually is an enterprise data warehouse. PING started building its data warehouse in 1989 with assistance from a vendor called Teradata . Over time, both the amount of data in the warehouse and variety of ways in which PING uses the data have grown significantly.

The concept of building a central platform for managing corporate data also has grown in popularity, with industry experts now calling it master data management.

At its most basic level, master data management involves taking data from disparate sources and putting it into a common format. This process—also called data rationalization or data cleansing—is a prerequisite for companies wanting to install business intelligence (BI) software or other tools that analyze operations.

That’s why BI software vendors like Teradata and SAS were among the first to develop master data management solutions. They were helping customers rationalize data stored in various operational systems—ERP, supply chain management, and others—for analysis purposes. Before long, companies like PING discovered that having clean data also makes it easier to create and implement new business processes.

There was a time when companies could get by with data residing on multiple systems, in different formats—even in separate business units. As markets become more competitive, companies must foster innovation in everything they do—from the way they design products to how they serve customers. Cross-functional data sharing becomes a necessity, which is triggering interest in master data management.

“New processes and applications require a lot of information to drive them,” says Bill Swanton, a VP with Boston-based AMR Research . “For example, to automate a business process, you need to know that, when you see this type of material, this is how we order it, this is how we store it; this is how we manage the inventory levels.”

In short, launching a new process quickly requires the ability to smoothly propagate data across the entire enterprise, and sometimes out into the supply chain.

Swanton says this enterprisewide data sharing can be especially challenging for processes that involve moving inventory—whether it’s finished goods, work-in-process, or raw materials. “Almost all information related to customers comes from the sales or customer service departments,” he says. “With materials, some of the data comes from procurement, some from engineering, some from the environmental health and safety office, and so on.”

Mainstream data management

The need for easy sharing of inventory or production data to execute business processes has caused a number of enterprise software vendors to develop their own master data management solutions. For instance, i2 Technologies , a supply chain management software vendor, has a solution dubbed Master Data Management.

Oracle , leading supplier of corporate databases and the No. 2 provider of ERP software, has developed architecture for building data hubs that support specific types of business processes, including the Customer Data hub, the Financial Consolidation data hub; and the Product Information hub.

“Data has always been stored in different functional areas—manufacturing, engineering, purchasing,” says Greg Thomas, technical director, solutions marketing for i2 Technologies, adding that i2’s Master Data Management solution helps companies rationalize overlapping and conflicting data in disparate systems. It also integrates transactional, analytical, and collaborative data to create an enterprise data warehouse, offering a single view of the business.

“You need this single view—one version of the truth—to do all the planning that drives a business, such as forecasting, supply, demand, transportation, distribution, and replenishment planning,” Thomas says.

Prior to MDM system implementation, i2 identifies three standard workflows that it calls enterprise processes, including new customer introduction, vendor data maintenance, and new item introduction. “We are seeing brand new business processes that typically involve different solutions, applications, or services, so they become composite applications composed of key information from different data points,” explains Thomas.

However, to run processes like these, you need a central view of what the data is going to be, especially the key data drivers—item data, customer data, and vendor data—as well as all of the supply chain data.

“Our data management technology allows companies to extend and add new business processes without complex overhead or additional software,” Thomas says. Users can get the data from the single data hub to roll out these business processes faster. For example, they can do a prototype process where they get the first version out quickly, do a test with users providing feedback, and then do another revision. And they can do this often, because the data is available so quickly.

According to Hardeep Gulati, senior director of product life-cycle management for Oracle, a sound MDM strategy ensures there is a good source of consistent product data available to manufacturing and sales during new product introduction.

“When you introduce a new product, it goes from product definition to engineering, then to the development organization, operations, purchasing, and so on,” explains Jerry Hill, VP and manufacturing industry consultant for Teradata, adding that this process is full of potential for data defects or missing information. “It could be something as simple as someone forgetting to order labels for the package,” he says, which delays new product introduction. What’s needed is data system integration—which is where master data management comes in, ensuring that the sequence occurs properly with the right business controls and expectations.

Master data management also is helpful during mergers and acquisitions, when disparate processes must meld into single processes serving the new, larger organization.

“If you are merging with or acquiring another company, you obviously see the advantages, whether it be filling a product gap you have, or adding new customers,” notes Mike Newkirk, industry marketing manager for SAS. “To take full advantage of this, you need to merge all of the data that exists in both organizations.” MDM can assist in such a data merge by providing an overall strategy that includes data migration, data quality, and data federation.

Now or later?

One challenge when using MDM for new process introductions is whether to have it in place already, then utilize it for the new processes; or wait to introduce an overall MDM strategy after the new process is in place. The answer is different for each organization and depends almost entirely on how senior management sees things.

“We find that companies introduce master data management when they are rolling out new business processes,” observes AMR’s Swanton. “No one will give you money just to improve your data. It has to be budgeted as part of a particular project.”

Swanton recommends selecting a large project that will have significant benefit to the company, and then building the MDM strategy around the process. Yet according to Teradata’s Hill, senior management may not initially understand MDM well enough to see the impact of delaying its adoption.

“When I was a manager in a semiconductor company, we would end up with late product introductions,” recalls Hill. “In other cases, we would get to the end of the new product introduction cycle, and the cost would end up being $150 a unit, compared to an original estimate of $100 a unit.” It wasn’t until much later, in looking back, that Hill realized the causes of most of these delays and cost overruns were problems with data. “If you can show management how these problems are caused by lack of a solid master data management strategy, they may be more willing to adopt one,” he states.

It’s a strategy, not a technology

When implementing MDM to support a new process introduction, it’s important to establish an enterprisewide data governance program that identifies and defines data elements and hierarchies, as well as policies and procedures for how master data will be maintained. It’s also important to assign data stewardship roles and responsibilities to ensure consistent definitions and resolve potential conflicts over data ownership.

“Success with master data management is much more dependent on people than technology,” emphasizes SAS’s Newkirk. “You can’t just buy the technology, install it, and assume you have master data management. The people and the process must come first, or the technology is doomed to failure. This requires senior-level commitment and support.”

Ultimately, the organization must find a way to get the individual business functions to surrender the concept of “data ownership” and get everyone comfortable with the idea of uniform, organization-owned data in which everyone can share.

One thing that eases the transition to master data management, according to Newkirk, is that users don’t have to give up applications they are accustomed to using.

“You also don’t have to try to integrate the different applications—which often results in failure,” says Newkirk. Rather than trying to integrate all the applications involved in a particular process, MDM simply integrates the data from all of those applications. That’s a much simpler process—and one that’s easier to manage over time.

Master data management (MDM) results

Value
What does MDM provide?
Business benefit

Source: Teradata

Data quality
Data validation and error checks to ensure input data is clean.
More accurate inventory, customer order commitments, and improved service levels.

Data integrity
Centralized data management for important entities; single portal for multiple users across multiple organizations.
More reliable reporting and analysis; reduced cycle time for introducing new products and vendors.

Data synchronization
Enables consistent data flow throughout the enterprise, and with trading partners.
Faster integration architecture for reduced latency.

Total cost of ownership
Standard workflows that can be quickly configured to business needs; most data management requirements are included; customization ability.
Lower deployment costs; lower maintenance costs; less risk.