[ About Info|Ed ] [ TiAC ] [ What's new at Info|Ed and TiAC ] [ TiAC White Paper Repository ] [ Information Architecure Resources ] [ White Paper Request ]

Data Administration & Repository Projects

Answers to Some Common Questions

© The Information Architects' Cooperative (TiAC) 1996

What is Business Data "Ownership" ?

The word "owner" has certain connotations that most people understand when it comes to tangible things like cars, houses, etc., but its meaning is less clear when it relates to data. The Corporation, the legal entity, actually owns the data that you and I manage on it's behalf in the course of conducting our work. When you and I leave this company, the data that we have managed and used will stay with the company, just like the furniture, phone, etc. that we have also used.

Every GM within the Corporation has the responsibility of ensuring the quality of the information needed by their staff in achieving the Business Unit's objectives. Individuals within the BU may have certain responsibilities delegated to them for subsets of the business' data (e.g.. materials data, employee data, financial data, etc.).

Accountability over the business' data is an active role, carried out throughout the life cycle of the data (from initial planning & creation to its disposition). It strives to ensure the quality (accuracy, completeness, integrity and timeliness) of the data and its continual evolution to meet changing business needs. It requires a sound knowledge of the business and of fundamental data management practices.

Therefore, it is not so much a question of ownership" as it is a matter of accountability; accountability for ensuring that the data needed by the business is managed cost effectively and supports the decision making processes at all levels of the business by all stakeholders. Some companies have used terms like custodian, partner, steward or guardian to encompass the responsibilities and accountabilities over the organization's data. The key consideration should be to pick a name that clearly conveys its meaning and its importance to the organization. Perhaps a better term to use is data stewardship. Based on the Webster's definition, a steward is '`a person, usually one of a group, put in charge of the affairs of a particular area, and acts as an administrator on behalf of others..."

Is Data Stewardship something new and why the recent interest in it ?

Accountability for data has always been there, even before the introduction of computers. Organizations tended to have rigorous procedures in place to manage files used by the business, including use of standard formats, check-in/check-out procedures, updating procedures and authorities, etc. With the introduction of computerized applications, each application managed it's own data files and introduced many of the same manual conventions and procedures. These early applications were generally used by a relatively small and closely working group of people (accounting, payroll, etc.). Accountability for the computerized data was relatively easy to manage.

Data Administration & Repository Project

As more sophisticated applications were developed and Data Bases were introduced, data became a greater shared resource used by a more disparate group of people and required specialized expertise or Data Base Administrators (DBAs) to manage it. Business accountability for the data became blurred. DBAs by default assumed part off the responsibilities, such as ensuring data integrity and establishing proper data security measures. However, these responsibilities tended to fade over time once the systems and databases were implemented. The situation was further exacerbated over time as data was copied to new platforms, out of the direct line of control of the original DBA.

With the proliferation of PCs and Local Area Network applications, it is not uncommon for 4 or 5 copies of the same data to exist, often in slightly different formats, updated at different frequencies and each managed according to the whims of the individual organizations that spent resources having the application developed. This results in an application oriented view for managing data. Data accountability is fragmented across many applications ail managing copies of the same or related information. It is increasingly difficult for the business to make timely decisions, because it is not clear that the data they have is the right data. Poor quality data impacts the bottom line: cost.

As the organization strives to move away from an application oriented environment, towards a shared data environment and a commitment to quality data, the accountability for improving the data situation rests on the shoulders of the business community whose business relies on the data, supported by an effective data administration function.

What are the responsibilities associated with Data Stewardship ?

Those entrusted with stewardship of the business' data, have certain responsibilities that they must carry out with respect to the data integral to their business, regardless of the medium it is stored in or the applications that manage it. These responsibilities include:

Does the sphere of Data Stewardship extend to an entire class of data, or is it limited to a smaller subset, even down to the data element level ?

This is probably the most difficult question to answer. Once the business processes are identified, and the processes that create the data of particular interest are known, the primary stakeholder and therefore the most likely steward for that data will come from the business process that CREATE the data. However there may be many other stakeholders and their views are important too. There needs to be a way to balance individual accountability and stakeholder involvement.

One approach would be to define classes of data based on data commonalities, usage commonalities, etc. and then establish `'common data interest groups" who would collectively provide the business knowledge about the data and their respective needs. The Data Steward would chair such a group and have accountability for the data throughout its life cycle. (e.g. Valve Common Data Interest Group). Common Data Interest Groups (CDIGs) could span organizations and/or business functions, or be totally self-contained. The objective is to assign specific accountability and maximize business flexibility and stakeholder involvement.

With the support of the Data Administration function, the Common Data Interest Groups would continually provide input to an evolving data plan or architecture, in support of the business objectives. The CDIGs would have accountability over the entire data class, regardless of which organizations or of how many information systems and/or users needed access to it

How does Data Administration support business in their stewardship over data ?

Data Administrators have specific data architecture skills that are essential to plan & maintain an evolving data map of the business and are used to properly plan relational data structures that:

In this technical capacity, Data Administration supports business groups in defining and capturing key business information about data and ensuring that a proper data architecture exists and is the focal point for all information systems initiatives in the organization. The information is cross referenced to the business functions and processes that use the data and is used to involve the right business stakeholders in any decision making with regards to the data. Data Administration maintains a global

Does all data need a Data Steward identified? Why is it important?

Yes, all data should have a data steward identified. This is especially important in a shared data environment, because the impact of poorly managed data will affect a large proportion of the business. Common sense of course needs to prevail. Less rigor is required for non-critical data, shared by a few users, versus critical data that a large number of business groups rely on. The degree of criticality, usage, shareability and value to the organization will be deciding factors

Business accountability for data is often identified in the course of defining new data entities, data elements or business rules, by information systems professionals during the requirements identification phase of an IS project.

Are there Data Management principles that the data stewards need to be aware of ?

There are many data management principles that can impact various phases of data management from data planning to ongoing operations. Good data management principles (Listed in Appendix A) serve to communicate clearly management's philosophy towards data accountability and form the basis for a sound Data Administration Policy.

Are Policies, Procedures and Guidelines necessary?

Yes. Strong Data Administration policies will ensure that data stewardship is part of "doing business" and is operationalized via the guidelines and procedures put in place. Everyone, from the shop floor level up to senior management have to recognize that they have a role to play in managing the company's data. Unless this is clear to everyone and enforced, as part of their job description or as part of their performance measure, poor quality data will continue to hamper the business.

Are data standards important and where are they needed?

If the business community is to assume responsibility for ensuring that business definitions for data are "good" then a clear set of guidelines and standards must exist to support their decision making. Data standards help ensure consistency in the way information about data is captured, documented and managed and to facilitate sharing of data and data management solutions. These can include data naming conventions, data validation standards as well as modeling conventions that will need to be established to facilitate communication of business needs using a common language.

How have other companies resolved Data Stewardship?

Many companies have recognized that proper data management starts with business taking responsibility over its data assets and putting proper procedures in place to ensure that proper data management practices are followed. Efforts vary from the non-scientific approach, where simple directories are used to identify primary contacts for classes of data, to sophisticated databases relating all information about the data down to the data element level. Again, a common sense approach needs to prevail.

The demands on the individuals accountable for the business' data will be many, especially during the early stages when roles and responsibilities are unfamiliar, and the data quality problems immense. The following skills will ensure that the individual is successful in this challenging role:

Where does a company begin in establishing proper accountability over data?

There is no one right way to go about establishing accountability for the organization's data assets. Establishment of proper data accountability starts by:

1. Identifying & widely communicating key data management principles
2. Identifying and communicating responsibilities associated with data stewardship
3. Identification of performance measures that recognize the value of data stewardship
4. Developing and issuing a strong Data Administration Policy
5. Establishing a Data Administration function to support business in carrying out their stewardship responsibilities
6. Creating an action plan to operationalize in small incremental steps.

Business input will be crucial in development of a business model, showing clearly the key business processes that need data, a high-level business data map (definitions, attributes and relationships), etc.