The global market for master data management (MDM) services and software will reach $10.4 billion by 2009, according
to predictions by Research firm IDC in Framingham, Mass.
Master data refers to vital business information that needs to be shared across multiple systems. Master data management is the combination of strategies and technologies to uniformly describe and present this data. The goal is to reuse common data to drive business value, according to analyst Andrew White at Stamford, Conn.-based Gartner Inc.
"Instead of handling information one project at a time, the CIO of the future is going to have to support the CEO's demand for [the company] to be more agile and nimble. That means bringing different groups of information together once and for all with a holistic strategy," with MDM as the centerpiece, White said.
Experts recommend following a series of steps to achieve master data management success.
- Attack pressing problems first
CIOs should zero in on particularly painful business issues that are well known throughout the organization.
"The most effective way is to identify the cost of errors that are directly attributable to inconsistent master data," said David Waddington, a vice president and research director at Ventana Research in San Mateo, Calif.
Having a unified view of data can help CIOs quantify the value of lost opportunities, missed sales or inefficient processes.
"One company told me that, prior to implementing master data management, it had the wrong information about case sizes on just a couple of products. It turned out they were running the equivalent of seven empty trucks across North America every week," said Bill Swanton, an master data management analyst at AMR Research Inc. in Boston.
- Win executive support
Showing such payoff makes it easier for top management to throw its weight behind an master data management rollout, as happened at Mentor Graphics Corp. in Wilsonville, Ore. The software maker is using master data management applications by Hyperion Solutions Corp. to solve several business issues, starting with customer accounts being incorrectly assigned to the wrong salespeople.
"Master data management is something that we in IT were driving, but there was strong sponsorship by the business, especially our chief accounting officer. He wanted this problem solved and he understood the benefits we were going to get from MDM," said Jan-Willem Beldman, Mentor's team lead of analytical applications and data quality, noting that changes to accounts now are made in minutes rather than weeks.
CIOs also should recognize the role of master data management practices in rolling out service-oriented architectures, in which information is presented as services across heterogeneous environments, White said.
"One of the primary things in your SOA strategy is developing new information architecture, and it's going to look exactly like MDM," White said.
- Gain consensus
Without consensus, experts contend master data management initiatives are doomed to fail.
"You've got to get agreement on which users are going to be responsible for creating and managing the information, and then you need to establish business rules around the data to ensure that it is being created correctly," Swanton said.
Henry Morris, an IDC analyst, advises CIOs to view master data management as a business process issue, not a technology issue.
"Find a single place where all master data should originate. It can be different places for each type of master data. And don't just think about the record of data, but which attributes need to be consistent across systems, and which attributes are of interest only to a particular local domain," Morris said.
- Delegate responsibility
Appointing certain people to be responsible for creating and managing specific types of master data could help. These "data stewards" should be provided with organizational support, incentive rewards and any technology tools necessary to do the job, according to Cliff Longman, chief technology officer at Kalido, an MDM software vendor in Burlington, Mass.
"The structure [you] set up should be culturally compatible with the organization," Longman said.
- Build your architecture
An master data management system consists of repositories for storing master data and any attendant metadata. Master data management software analyzes information in various systems to make sure it is accurate and can perform the intended tasks. Also needed is a workflow engine to choreograph the lifecycle of data, as well as applications for customer data integration, product information management and others.
"Every company needs to have a master data management strategy. How they solve that problem will depend largely on two things: the relative use of off-the-shelf vs. custom software, and how consolidated their application landscape is," Swanton said.
When constructing a data model, White advises choosing vendors with deep experience in applicable taxonomies or industry expertise. The important thing, White said, is to "make it your data model, not a vendor's data model, because everything [in your master data] will then be matched to it."
- Take a long view of data
Improving the management of master data is no quick fix. Users and experts say it's wise to take a long-range view and implement master data management projects incrementally. It could take many months, if not years, before you arrive at the finish line.
Says Belden of Mentor Graphics: "Attack each problem one at a time. Get good executive sponsorship. And don't underestimate the time it will take to get everyone pointed in the same direction."
Garry Kranz is a freelance business and technology writer in Richmond, Va. He can be reached at email@example.com.