Managing information as an asset requires knowledgeable business leaders.
by Jonathan G. Geiger
Most organizations understand what it takes to manage assets such as money,
facilities, employees and inventory. However, when it comes to one
asset—data—companies often fail to deploy the responsibilities and
accountabilities needed to effectively manage it. A common result is that data
availability, accuracy, timeliness, protection and accessibility issues abound.
Additionally, when employees aren't certain who is responsible for data
management, they may point fingers at each other and scramble to address the
symptoms without truly resolving the problems.
Every other business asset has an organizational accountability assigned to
manage it. Data should be no exception. To truly manage data assets, an
enterprise needs to establish a cross-functional governance committee and
designate data stewards within its business organizations. The governance
committee should promote the importance of managing data as an asset, establish
policies and stewardship responsibilities, and enforce compliance. Stewards
should have the knowledge, responsibility and authority to describe, establish,
declare and enforce business rules about data. This ongoing role is ultimately
responsible for establishing quality expectations and ensuring that they are
|Data steward requirements
Business knowledge: Data stewards must understand the business
direction, processes, rules, requirements and deficiencies.
Business-area respect: They need to influence business
decisions and gain business-area commitments.
Analysis: When faced with multiple options, they must examine
situations from many angles.
Facilitation and negotiation: They must facilitate the
proponents of conflicting viewpoints to arrive at a mutually satisfactory
Communication: Stewards need to effectively convey the
business rules and definitions and promote them with the business areas as
Data management responsibilities
The chief financial officer (CFO) is in charge of planning, acquiring,
maintaining and distributing financial assets; the personnel director has
similar responsibilities for human resources, etc. But who assumes these
responsibilities for data? For example, asset planning includes establishing
how the asset is categorized and organized as well as forecasting its needs.
For financial assets, the CFO develops the chart of accounts and (with input
from other areas) projects revenues and expenses. For data, the equivalent of
the chart of accounts is the business data model. Data requirements are
analogous to the revenue and expense projections, and the metadata management
system is functionally similar to the general ledger system.
While the chief information officer is recognized as having authority
associated with managing the data asset, other business areas also claim
responsibilities, resulting in overlaps (creating tension, redundancy and
inconsistencies) and gaps (creating deficiencies). This problem can be resolved
by separating the responsibilities into custodianship and stewardship.
Custodians are responsible for the "buckets" that store the data, and for the
systems that process it; stewards are responsible for the "content"—the data
IT is the custodian of an organization's data assets. This department gathers
business requirements, develops the application systems, operates these systems
and supports their users. Further, it manages the infrastructure on which these
applications operate and ensures that the data is properly stored and
protected. Data protection, for example, might be accomplished by providing a
security system and periodic backups. To perform this role, IT establishes
processes such as the system's development methodology, facilitates the
development of the business data model, and sets security administration and
Business representatives (stewards) must retain responsibility for the data
content. Stewards define the requirements, verify that systems meet them and
use the systems. The business representatives also have a data protection
responsibility. This includes determining who can retrieve what data and
enforcing access restrictions.
The table below contrasts some of IT's custodianship duties with
stewardship responsibilities performed by business representatives.
Data steward qualifications and deployment options
A data steward needs to be a seasoned analyst who understands the business and
data management concepts and is able to recommend and gain reasonable
compromises that enhance the value of an organization's data assets. Specific
competencies include business knowledge, business-area respect, data management
knowledge, analysis, facilitation and negotiation, and communication.
The data stewardship function may be established either formally or informally.
The formal approach entails creation of the aforementioned cross-functional
governance body and appointment of stewards for the data areas being addressed.
The stewards would have job descriptions that delineate their responsibilities.
This approach is viable in mature organizations that truly embrace data as an
A pragmatic and informal approach to get data stewardship started is to use
subject matter experts (SMEs) from the business organizations to provide at
least some of the stewardship functions without the formal designation. In a
data warehouse project, for example, business input on the enterprise
definitions and business rules are obtained from SMEs. These experts contribute
to the business (and data warehouse) data models and represent the business
areas in defining the enterprise views. Successes and shortcomings of the
informal stewardship role should be documented as each project proceeds, and,
at some point, the case should be built to formalize the function.
The data steward is a business representative who is ultimately accountable for
determining, describing and enforcing the business rules and definitions for
data. For long-term success, this effort should be supported by the
establishment of a cross-functional governance body to ensure that data is
managed as an asset. When a stewardship function is needed for a data warehouse
and a formal program is not in place, a bottom-up approach can bring short-term
success and provide support for building a viable data stewardship function.
|Custodianship versus stewardship responsibilities
Ultimate success depends on cooperation among data stewards as well as a strong
ongoing partnership between stewards and custodians.
|Create the business data model.
Facilitate creation of the business data model, maintain it and apply it for
data management database design.
|Contribute to the business data model.
Provide the business rules, definitions, etc.
|Gather business requirements.
Interview data stewards and other business representatives to define the
business requirements. IT team members gathering the requirements must apply
interviewing and analytic skills to ensure that they ferret out the
requirements and translate these into the information required to understand
what needs to be designed and built.
|Provide business requirements.
Work with IT analysts during the system development and provide them with the
|Establish data access restriction facilities.
Install security mechanisms that enable data access restrictions.
|Determine who can access what data.
Input information on who has access capabilities into the data access
|Deliver quality data with supporting audits and controls, and resolve
Build systems that meet business expectations and provide evidence (through
audits and controls) that the data has been processed correctly. Errors
disclosed through audits and controls indicate data processing deficiencies and
must be addressed by IT.
|Establish quality expectations and ensure compliance.
Ensure that realistic quality expectations are consistent with business
processes. In addition, ensure that the data conforms to the quality
|Incorporate changes as needed.
Continuously communicate with the data stewards and other business
representatives to keep informed about business changes that impact the systems
and data. When these occur, gather detailed requirements and incorporate the
changes to the existing environment.
|Keep IT informed of business changes that impact data.
As the business environment evolves, changes are needed in systems and data to
ensure that they continue to support the business. Information about these
changes must be transmitted to IT on a timely basis to ensure that the data
asset can be appropriately leveraged.
|Critical success factors
These activities must be performed well to survive in the short run and thrive
in the long run:
1. The enterprise must recognize data as a vital asset.
The stewardship function requires a commitment to truly manage data. Without
this recognition, it is unlikely that the function would even be created.
2. An active cross-functional governance body must be in place.
Data is cross-functional in nature, and several departments might claim
responsibility for certain data. A cross-functional governance body is needed
to divide responsibilities and to resolve conflicts. In addition, the
governance body determines overall data stewardship policies (e.g., the
organizational accountability and authority of the stewards) and establishes
priorities. The governance body must also promote the importance of data as an
asset and support the data stewards in the performance of their duties.
3. Data stewards must work together.
Unlike business processes, which tend to be aligned with the organizational
structure, data crosses departmental lines. For example, the sales, marketing
and customer service departments may all claim responsibility for customer
data. While some of this data might clearly fall within the purview of one of
these departments, all three have a claim on basic consumer information such as
name, address, selected demographics, and the processes and systems for
acquiring and managing this data. Unless a steward with ultimate
decision-making authority for this data can be appointed, stewards from these
areas must work together to establish a consistent set of definitions and
4. A business data model must be created, maintained and applied.
This model describes the business concepts, events, places and people of
interest to the organization and the rules governing them. There is only one
such model in an enterprise, and its existence and application ensure
consistency and reconcilability among multiple physical databases that may
store the same data.
5. The assigned data stewards must understand their role, be qualified to
perform it and believe in its importance.
Data stewardship is not an easy job. In addition to possessing the needed
qualifications, the person performing this role has to recognize its importance
to the organization. A steward must be willing to tackle difficult issues such
as resolving conflicts among seasoned business and IT professionals to arrive
at a solution that correctly describes the business rules concerning a set of
Jonathan G. Geiger is executive vice president of Intelligent Solutions, Inc.
In his consulting, he specializes in data warehousing, customer relationship
management and data management. He has written more than 60 articles on these
topics, is a frequent speaker at national and international conferences and is
a co-author of three books.
Teradata Magazine-September 2008