THE DATA ADMINISTRATION NEWSLETTER – TDAN.com
ROBERT S. SEINER – PUBLISHER

Published in TDAN.com October 2002

Introduction

To effectively manage your company’s information assets – data, content & knowledge – it is important to identify and record information about the individuals that are accountable for the quality of these assets. Accountability, especially formal (and communicated) accountability, has been known to drive behavior.

Accountability for what … you may ask. In the field of information asset management … Accountable for defining and making certain that information assets are defined according to the way the business intends to use them; Accountable for making certain information assets are re-used rather than re-created; Accountable for making certain the data assets are of high quality both in terms of design and accuracy. And the list of accountabilities can go on ...

So the next questions may be … How do you make people in your company accountable for the information assets? How do you implement stewardship, data governance, insert word de jour, basically accountability for information assets? To these questions, there are no simple answers. Every company has a different perspective on what the right level of accountability means. Every company needs to handle stewardship or governance differently.

To start - accountability already runs rampant in most organizations. So many people are accountable for so many of the same and different things that it has become a “sea” of accountability. And nobody is absolutely certain of who is accountable for what. In fact, accountability becomes impossible to manage if left to its own devices. That is why there is a need to formalize stewardship.

When something is three-dimensional it is real. Merriam-Webster’s College Dictionary calls three-dimensions: describing or being described in well-rounded completeness. This article describes the creation of a formal stewardship solution in terms of being 3-D. Or should I say, in terms of three Ds.

This article will focus on accountability for data assets. The 3-Ds in this article refer to “De facto”, “Discipline”, and “Database”. If you can impact all of the 3-Ds of stewardship in your company, you will start moving in the direction of formal accountability for data assets.

The First D – “de facto”

Stewards, those individuals that have accountability for data assets, already exist in your organization. They are, most likely, “de facto” stewards by nature. You would have never gotten as far as you have without some level of accountability. The problem with this accountability is that it is informal. There are so many people accountable for defining data, creating data, managing data, … and often these stewards are defining, creating, and managing data that also already exists in another place (i.e. system) in your company.

One insurance company I worked for had 17 “master” provider files. Someone was accountable for how the data was defined for each of these “master” files. Often the same data was being defined just a little bit differently, but different enough to make it very difficult to integrate this data. Does this sound like what has happened in your company? So many people are accountable in a very ad-hoc sort of way.

Stewardship involves knowing how data is going to be used by your company. The same pieces of data are being used all over your company. As in the case of the “master” provider files, the providers account identifier was just different enough that it cost millions of dollars to scrub and match providers and move toward a single “master” file. Now multiply that single data field problem, a problem of definition of data – a problem of creation of data – a problem of usage of data, a thousand times and that may indicate why companies have the data problems that they do.

Data comes from up stream and flows down stream. Companies typically have “de facto” stewards at every bend of the stream. “de facto” stewards are not formally assigned the responsibility but rather they take on the responsibility of managing the data per their own needs and interests. This causes problems ...

1. If the information about “who the stewards are” and “what data they steward”, is not recorded anywhere (see the Third D for database), data will continue to be duplicated, data integration will continue to cost small fortunes, and people will always work under the premise of “best guess” when it comes to defining how data will be shared and leveraged as a competitive advantage.
2. Stewards at every bend cause the same data to be defined again and again. And perhaps defined differently again and again. By recording who is responsible for what, the company has a way of tracking how data is used throughout the company. By recording this information the company has the ability to bring together all of the stakeholders to create a common definition of how the data is used throughout the company.

It is okay to have “de facto” stewards in your company and long as you know who these individuals are. Once you know who the “de facto” stewards are, your company can use this stewardship knowledge base to build a roadmap to your company’s data. The recorded steward information can also be used to bring “de facto” stewards together to identify and resolve discrepancies in how similar data is used, created, updated, ... across the enterprise.

Ultimately, the identification of the “de facto” stewards can help your company identify a single person or group of persons (committee) who have the enterprise’s best interests at heart and have the final say on how the data is used and created.

The Second D – “Discipline”

There are several disciplines involved in the implementation of a stewardship program. These disciplines include ...

• Identifying and recording information about people that are accountable for managing the data.

• Using this recorded information in a way that it can be used "pro-actively" to avoid conflicts and data integration issues and "reactively" to directly impact the resolution of known business data integration issues.

• Involving the right stewards at the right points in the data and project lifecycle.

• Formalizing stewardship through information policy and performance evaluation.

In many situations, companies are not disciplined in how they record information about people and their specific accountabilities to manage data assets. To implement a stewardship program, it is imperative that the relationship between people and data be recorded in a database (The Third D) such that this information can be used by others in the organization.

A second discipline involves the use of the steward information "pro-actively" to prevent data integration issues from occurring and "reactively" bringing to resolution known data integration issues. Once the steward information is recorded in a database, people need to get used to using the steward knowledge-base to learn about how new data or changes to data may impact data and data processes downstream (as data is passed from application to application). The same knowledge-base can be used to engage all of the right people while reacting to known data integration issues.

If a company follows a structured development methodology, it is important to recognize the need for discipline in utilizing stewards during specific tasks in that methodology. For example: Data definers must be involved in the earliest stage of requirements gathering while data creators must be involved in the database design. Making certain that stewards are utilized the appropriate points in the methodology provides the working basis for both "proactive" and "reactive" stewardship.

The last discipline to be mentioned here looks at stewardship from an executive management perspective and includes the definition of the stewardship program in the enterprise information policy. Information policies typically state the "data is a corporate asset" and that the "management of the data is the accountability of people who represent the company and not their own interests". To further enforce stewardship and executive support of such a stewardship program, some companies are writing accountabilities for data assets into directors and managers job descriptions. In these situations, this description is the basis for these management individual's performance appraisals.

The Third D – “Database”

The steward database (or repository) plays a vital role in the success of a stewardship program. The steward database contains data on the relationships between people and the data; specifically the people that are accountable for defining how the data will be used, how the data should be created, updated, and managed.

Steward data is meta-data in the truest sense of the word. Data about data; only in this sense it is data about the people who manage the data. Steward meta-data can be stored in an enterprise meta-data repository or it can be managed in a stand-alone database.

Consideration must be given to steward meta-data distribution. Up-to-date steward meta-data will need to be made easily accessible to senior management, directors, managers, project teams, and individual team members in order for the stewardship program to be effective.

Consideration must also be given to the ability to manually enter and alter steward meta-data directly in the database. Whereas, most meta-data in an enterprise meta-data repository comes from another tool (data modeling tool, DBMS catalog, ETL tool, query tool), stewardship meta-data must be manually researched and manually entered into the repository.

Once steward meta-data is entered into the stewardship database, it is imperative that this information is kept up to date. This is an additional discipline (The Second D) that requires attention – someone must be the steward of the steward meta-data.

When building a new stewardship database or when extending an existing repository, companies should consider using existing “people” data that exists in HR or ERP systems as the “people” portion of the meta-data populating the stewardship database. This connection to HR systems will be beneficial if your company decides to automate the identification of steward meta-data that changed due to personnel moves (tracked in the HR systems).

Summary

Stewardship fully implemented brings out the best qualities of data administration, data re-use, efficiency in data integration and application development. Getting to that point, where data is managed with precision, is a long battle. Understanding how data is used across the company and knowing who in the company is accountable and responsible for managing this data are big steps toward improvement of enterprise data management.

The simple-to-remember 3-Ds of stewardship start with the identifying “de facto” stewards, proceeds into building “discipline” into how a steward knowledge-base is used, and ends by building a “database” of stewardship information.

Stewardship is not a short journey and involves long-term changes at the heart of how data and applications are developed. But … before any company arrives at a sound stewardship program, they should begin by focusing on the 3-Ds. You won’t successfully implement stewardship unless you follow a disciplined approach and record stewardship information as a course of action. The 3-Ds will be the backbone and the engine of any successful stewardship program.