Now that master data management (MDM) is on the short-list for many CIOs, it is
worthwhile to consider how MDM projects are spawning in organizations and how
an organization can evaluate their potential for success. Recent surveys by some
industry analysts recommend what the software mega-vendors such as IBM, SAP, and
Oracle have been hyping -- that CIOs must look at the MDM effort holistically
and embark upon an enterprise-wide MDM initiative. Today, MDM is a seductive vision
that is sold primarily to the CIO office. However, the mainstream success of MDM
depends on enrolling business as the sponsor or the master of such
initiatives, because master data is inherently tied to solving business problems.
Rather than falling victim to mega-vendor promises, CIOs are now opting for a
more prudent approach that enables them to start small with a business-sponsored
project to demonstrate the high return on investment (ROI) that can be realized
using MDM technology, and then evolve that technology into an enterprise-wide
MDM platform over time.
Data Governance in Infancy
The MDM vision being sold to the CIO paints a picture where all information
silos within the organization are dissolved and the data is set free to flow
among systems in real-time; data that is accurately unified with other data
and is transmitted securely and viewed discretely by each business user at their
desired frequency and latency. At a loftier level, the IT vision being presented
is one that supports a set of universal definitions of all core business data
entities or master data that are shared across all business processes
and systems.
This vision is certainly alluring but also appropriate in order to tear down
decades of legacy data silos. However, the first task CIOs face in realizing
this vision is addressing data governance issues. Specifically, organizations
must determine who defines the customer and standard product description; who
knows of the correct relationship between customer and organization; who resolves
the conflicts among these data sources, and more importantly, who owns the data?
With all these considerations, it is not surprising that settling data governance
policies and practices is an enormous, corporate-wide undertaking that has the
potential for political strife.
Ironically, the critical assumption underlying the MDM debate today is that
IT can manage master data. This implies that IT can anticipate, define, and
standardize on a complete MDM technology stack in advance of the negotiated
outcome of data governance issues between business and IT. The assumption is
flawed and may prove to be fatal for the entire master data management category.
Realistically, it could take three to five years at most organizations for the
critical data governance issues and organizational priorities to be fully resolved,
and for data governance policies and processes to be implemented.
-1-