• SELECT SITE CURRENCY
Select a currency for use throughout the site
Understanding Master Data Management
Ovum, May 2006, Pages: 27
Master data management (MDM) is not a new concept – the problem of managing master data has been with us since the beginning of computerised business automation. The topic is currently experiencing a major resurgence in interest because of a renewed desire to exploit data for better business effect, and because MDM is inextricably linked to the current interest in and the drive towards service-oriented architectures (SOAs). With this renewed attention and corresponding hype there is also an equal dose of confusion and misunderstanding surrounding the basic principles of MDM. Most of this confusion stems from inconsistent definitions and conflicting messages from vendors – each vying for that illusive market and mindshare – about what master data actually is, how it should be managed and what technologies should be employed to solve the master data problem.
This report details the basic principles involved in managing master data. It provides a comprehensive definition of master data, details how master data should be managed, and defines the functional scope and main components of a MDM solution. It also discusses the architectural approaches that need to be considered as part of an MDM implementation.
What is master data management?
Scope of this report
The MDM problem
The broken promise of packaged business applications
Increasing heterogeneity = disparate and siloed master data
Data proliferation = lost business opportunities
Why is master data such a problem?
The MDM promise
Managing master data
Master data information domains
Master data characteristics
Technical architectures for MDM
Master data record
Three approaches to MDM
Reference application MDM
MDM and SOA
SOA: from BPM to MDM
The split personalities of MDM
Application versus data integration
Packaged versus DIY approach
From data to applications
Beware of silos between solutions
Operational versus analytical data
Structured versus unstructured data