LAS VEGAS -- Large organizations, with all of their complexity and varying business units, may never achieve complete...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
master data management (MDM), but speakers and attendees at last week's Data Warehousing Institute (TDWI) conference say there are plenty of reasons to make the effort anyway.
Speaking to conference attendees, Cliff Longman, chief technology officer of Kalido, a Burlington, Mass.-based MDM software vendor, said that the benefits of MDM, a new discipline that combines software with data governance techniques, are plentiful and have the potential to impact all aspects of an organization from finance and operations to product development and production.
Longman admits, however, that especially for large organizations, getting to true master data management can be a lifelong process.
"You're all going to be here in 10 years' time, talking about how you've now gotten 30% of the way through the task, and feeling good about it and having gotten value out of it," Longman quipped. "And hopefully [you will be] proving that you have added good value from it in the business."
What is MDM?
MDM has been much hyped recently because of its promise to solve the old problem of achieving greater data consistency across an organization.
Many companies are attracted to the idea of MDM because they understand that true data consistency leads to things like more accurate reporting and business intelligence (BI) and ultimately reduces errors, saves time and improves the bottom line.
The technology itself acts as middleware and gives systems a common place to look for approved data definitions. But MDM experts like Longman say that true MDM must go beyond software to include data governance best practices and a tight working relationship between business and IT.
"This is not just an IT thing," Longman said. "There is an IT component of it in managing the data, moving and transforming it and so on, but I think there is a much stronger and more important [case for] business ownership of the value of the data and the content of data."
Longman said that MDM should be looked at as an enterprise-wide discipline, but that it's important to start small.
"The implications are that you have to have a flexible approach to doing this and a flexible technology to do it with," he said.
Benefits of MDM
MDM has the potential to impact all aspects of an organization, and companies like Unilever and British Petroleum (BP) are already saving hundreds of millions of dollars as a result of the technology, Longman said.
Other benefits of MDM, according to Longman, include the following:
- In finance and operations, MDM can lead to more accurate reporting and BI, more efficient planning and budgeting and enhanced ability for regulatory compliance.
- In sales and marketing, solid MDM can provide a "single view of the customer," leading to greater customer satisfaction.
- In procurement and supply chain management, MDM could mean reduced delays in shipment and fewer errors in item coding.
- For product developers, MDM means faster product launches and a better handle on customer feedback
- For IT, MDM could lead to improved productivity of standards-based application development.
Users seek one truth
At Progress Energy, the large number of independent business units and the complex ways those units interact means that getting information to sync up isn't always easy, said conference attendee Bruce Novell, a business application architect with the Raleigh, North Carolina-based power generation firm.
Novell explained that information is often shared informally at his company, through e-mail, spreadsheets and shared local area networks. With all of that information shuffling going on, he says that data inconsistencies can and often do show up.
While Progress Energy has had data warehouses for years for financial reporting purposes, the firm is now looking to MDM as a way to gain consistency -- and ultimately better analysis and reporting -- across the entire organization.
"We have many copies of the truth, and we're interested in the value proposition of, in many cases, eliminating duplication and streamlining processes, but also getting the benefits of consistency, whether it [is through] reporting or integration," Novell said. "We're just starting to look at it at the enterprise level."
At this point, Novell doesn't think that Progress Energy will ever have just one MDM system in place. And in some areas, he said, it's more important to reach the goal of MDM than others. That's why he says it's key for IT and business to get together and prioritize before embarking on any MDM plan.
"It probably will never be perfect, because no one will ever have the justification to fix it all," Novell said. "If I have three phone numbers and my home phone number is incorrect, is it worth any effort at all to fix that? There are lots of examples where the marginal cost is not worth the effort."
William Foster, a senior business data strategist with IBM who works on IBM's internal information management projects, says his firm, which hosts systems for customer companies the world over, has begun taking a new approach to managing clients' information that involves transitioning the management of source data from the customer side to a centralized database.
IBM customer data -- which had previously been managed through various customer master record systems -- is now being rationalized and fed through an interface into a centralized master data environment, Foster exclaimed.
Eventually, Foster added, that central database is going to become the master file -- the place where changes to source information are applied and then disseminated out.
"Ultimately we're going to get rid of all those individual customer master record systems throughout the world and just have that one central [hub]," Foster said.
Foster said this transition is still in progress and has been difficult at times.
"One of the biggest challenges was finding out what data needed to be rationalized and what data needed to be actually changed in the sources so that when it all came together in one database, it actually fit and made sense," he said.
This article originally appeared on SearchOracle.com.