A Passion for Metadata – An Interview with Todd Stephens of BellSouth

R. Todd Stephens is the Director of the Metadata Services Group for the BellSouth Corporation, a role he established in 1999.  One of the things that makes Todd particularly interesting as the subject of this discussion is that his laid-back personality and easy sense of humor disguise the “driven” man within – he’s a guy with both business acumen and serious technical smarts.  AND he understands how to make himself and his department indispensable – he’s constantly increasing the value of his department by taking on new responsibilities and new technologies.  At this stage his group is involved in architectural efforts that effect the development and implementation of database metadata, data transformation, components, XML, content, documentation, web services, messaging, metrics, interfaces, and the Enterprise Information Portal.  Todd is setting an example for the next generation of corporate metadata management – he’s business-driven, financially accountable, technically savvy and most of all he’s wildly enthusiastic about what metadata can do for the organization. 

Todd has multiple technical degrees, plus an MBA, and is pursuing his Ph.D. in Information Systems at Nova Southeastern University. The majority of his research is focused on Metadata Reuse, Semantic Zooming, Trust Usability and Repository Frameworks. He has been awarded six U.S. pending patents in the field of Metadata. And in the past couple of weeks his wife gave birth to their second child.  I sincerely appreciate him talking to me during this hectic time.

Tony Shaw, Wilshire Conferences (Wilshire): Todd, how did your Metadata Services Group get started at BellSouth, and how has it evolved since then?

R. Todd Stephens: In February of 2000, the Metadata Services Group (MSG) was formed to research, design and implement an enterprise metadata strategy for BellSouth.  The MSG is in the final stages of implementing a comprehensive, integrated solution to provide enterprise Metadata to the organization.  This initial directive has not changed over the past several years.  However, the method and scope of delivery has changed and will continue to evolve as we move forward.

Initially, the MSG was challenged to implement the metadata repository over the enterprise databases.  We were able to setup the infrastructure and application very quickly in order to fulfill this task.  The enterprise data models, physical databases, legacy sources and transformation information were loaded into the repository and updated on a regular basis.  However, it didn’t take long for the customers to begin to ask questions that this initial effort couldn’t answer.  Questions like:

• Where is the design document that provides the detail mapping information?

• Which system and interfaces are associated with these legacy applications?

• How much data is loaded and what are the error rates?

• How can I gain access to this enterprise information?

• Is this metadata information available for the middleware platform?

We started to see that we needed a lot more information than what traditional metadata and data transformation could provide.  The team pulled together and decided that if we are going to be in the repository business then we are going to provide all of the information required by a true enterprise effort.  Our organization is focused on delivering metadata management at the enterprise level which allows us build a wide variety of repositories that go beyond the traditional metadata implementation.  The repositories contain information on database metadata, data movement, components, documents, content, XML artifacts, web services, systems, interfaces and many other content categories.  

Wilshire: Let’s get into the specifics of some of those initiatives shortly, but can you tell me how did BellSouth decide it wanted to be doing these things? How did you get involved personally?

Stephens: Enterprise Application Integration (EAI) represents a significant evolution in integrating disparate islands of automation within the corporate enterprise. From BellSouth’s EAI effort emerged a collection of enterprise service organizations that provide services across the corporation; the Metadata Services Group is one of them.

If the EAI effort was the spark that started the metadata fire, then integrating metadata management into the data architecture was the fuel. One of the first questions I ask people that come up to me to discuss their metadata implementation is “Do you have a corporate architecture environment?” And “Is the data architecture governed at the corporate level?” If you can get the principles, policies, guidelines and standards raised to the architecture level then projects must file variances on why they won’t provide the metadata information required by the data architecture. This requirement not only applies to the internal organizations but also to external vendors. When application providers want to sell us a packaged solution, they must answer these questions in writing:

• How will you represent the data? The models? How frequent will your updates be?
• Do you have the detailed metadata information and how will you export or import the information?
• Is your API layer documented? UML Models?
• Are your data edits and data quality processes documented? Where?

Interesting enough, I don’t really have a background in data; at least not from a data steward or modeling point of view. I came through the programming ranks in the IBM AS/400 world. In late 1999, I found myself in a hardware architect role, which my skills were not well-suited for. On a particular assignment, I worked with a gentleman that had just been assigned a role that he wasn’t exactly thrilled with. That role was to lead an enterprise effort in the implementation of metadata management. As we chatted, we discovered that my job was perfect for him and his job sounded exciting to me. We simply switched jobs and the rest, as they say, is history.

Wilshire: I know you don’t restrict your definition of metadata to “data about data.” How do you define “Enterprise Metadata?”

Stephens: I look at Enterprise Metadata as the structured, semi-structured, and unstructured information that describes the characteristics of a resource or asset. Metadata is about knowledge, which is the ability to turn information and data into effective action. Have you ever wondered why data management professionals spend so much time trying to convince their management that data is an asset, but then not expand the utility of metadata beyond data?

We see data as one of the many technological assets a corporation has implemented and then we simply expand that utility to the other assets of the corporation. Take impact analysis for example, we all want to know that if we change field A then we should also modify fields B and C. However, what about document X, component Y and the message structure Z? They need to be modified as well. Metadata cannot survive at the enterprise level by just focusing on data. The audience is simply too narrow.

Wilshire: Let’s hear some more about your “Metadata Services Group.” Can you elaborate please on how the Group operates, what its objectives are, its organizing principles, etc.?

Stephens: The Metadata Services Group set out to re-define metadata management within BellSouth. We are a team of eight dedicated individuals with a passion for metadata. Each member has their own set of unique skills and experience; some on the technical side and others on the business side. Together, they have created an organization that is known through out the company and the industry. There are so many reasons why enterprise metadata management should fail within an organization: no funding, no integrated toolset, no business buy-in, no requirements, no defined users, etc. But once you start falling into the trap of trying to explain failure, failure is inevitable. Many of the repositories emerged from roadblocks put up by different organizations. Each of the roadblocks was met with new products and services; they simply ran out of excuses for not implementing metadata. I can’t say enough about the talented group of individuals we have working in this organization. Without them, none of our success would have been possible.

We have five basic principles that we operate under and we refer to these as the MSG “High Fives”:

1. Simplicity: Keep it simple so that anyone can understand the value.
2. Preparation: Always, Always, Always be prepared for anything.
3. Perfection: Strive to do your best, each and every day.
4. Pride: Have pride in your team, your work and yourself.
5. Customer Service: Serving others is all that we do.

Wilshire: I have to pick up on that comment for a moment – how do you get people to have a “passion for metadata?” How do you make it glamorous and desirable, or even sexy?

Stephens: Let me correct the statement on the “passion for metadata”; I don’t necessarily look for passion in metadata but rather passion for a “piece” of the puzzle. Some people on the team have a passion for quality. They will relentlessly pursue the quality of metadata in each and every load we have. Others have a passion for technology, and they will deliver the best applications because they are driven by the desire to build the very best. This is a wonderful situation to find your self in because I don’t have to worry about things like data quality or the technology. I already have someone that is far more concerned than I would be on this subject. This allows all of us to focus on what we love, what we believe in, and what we have a passion for.

The people on the team that don’t last more than six months are the ones that don’t develop a passion for some aspect of the department. Metadata is not a social club and I can promise you that if every person on the team had my passion for metadata, it would be one obnoxious environment. In many ways the staff is motivated by achievement and it is irrelevant whether that achievement is personal or professional. People instinctively want to part of something special. They want to believe that their work matters and will make a difference in the direction of the company. Simply put, Metadata Matters!

Wilshire: That’s starting to lead into my next question then…how do you sell yourselves within your organization?

Stephens: Selling, marketing and branding metadata management is about creating a perception of value. Keep in mind that metadata and data management is going to have a perception, the question is will that perception be based on a planned design or some default. We are constantly looking for ways to keep our name and what we stand for in front of people. Last year alone, we did 14 lunch and learns, 4 town halls, 9 conferences, published 6 articles, 12 white papers, 2 newsletters, 5 architecture presentations and filed 4 new patents. Not to mention, we provide Intranet publishing services for 28 different organizations. And, whose logo is at the bottom of every page? Metadata Services Group, of course.

Keep in mind, generating all of this exposure would be the death-nail if we didn’t have the product, service, knowledge and ability to back it up. It only takes a couple of issues or problems to destroy the perceptions we have created. We have fun with the concept of branding our group, but this is serious stuff and an enormous amount of thought goes into how we want our organization to be perceived. We don’t let anything go to chance. Just as quality cannot be a random event, neither can the perception of the organization.

Wilshire: I’ve heard that you don’t have any problems getting your metadata budget renewed each year? Is that true?

Stephens: BellSouth has never wavered in the investment of metadata and data management. They have provided me with all of the resources that I have asked for over the past three years and continue that support us for a fourth year. Despite the majority opinion, getting budget dollars for metadata is the easy part. Defining, presenting and demonstrating a return on the investment is the challenge. When you approach an organization with a solid business plan, demonstrated value and enthusiasm, they will gladly provide you the means necessary to deliver the proposed solution. They are investing in you, and your ideas. Not much different than people getting venture capital to start a new business. Executive management wants to see that you have thought about the environment, planned appropriately and can deliver a return on their investment. Please don’t lose my point here; the company is not just investing in data management they are investing in you.

Wilshire: You’ll be providing these and other insights into your success story during your tutorial at the Wilshire Meta-Data Conference and DAMA Symposium in Orlando on April 28. What other “teasers” can you give us about what you’ll be talking about down there?

Stephens: Sure, how about these:

• You will learn how we returned 80% of the initial metadata investment in 3 minutes.
• You will understand why metadata must be raised to the architecture level in order to be implemented across business units.
• You will learn what we mean by “Metacide”: The killing of a metadata project.
• You will learn how to use the Dublin Core standard.
• You will understand why data architecture is so hard and perhaps why it is not.

That’s just a few of the items. The best sessions that I have had the privilege to attend at the DAMA/Wilshire conference are the case studies. They reveal more information and experience than a dozen high level overviews. In fact, a lot of our success can be tied back to the UPS case study that I attended at the 2000 DAMA/Wilshire conference held in Washington, D.C. That one hour session didn’t provide all of the information I needed, but it did open my eyes to the possibilities of metadata.

Wilshire: You’re also an early adopter of new technologies. What’s been your experience with XML so far? Where do you see its usefulness to data management?

Stephens: Yes, we have jumped all over the XML band wagon. I have two points of view when it comes to the potential of XML. First, XML is one of the first times I can remember where the hyped technology is truly about data and metadata. Whether you agree or disagree with the hype, you can’t ignore the impact XML has had on our industry. XML has truly moved the concepts of metadata and data management to the forefront of the industry. Can I predict the future of XML? Not any better than you can but what I can do is prepare for its integration into the corporate framework. They way we are doing this is by providing BellSouth with a local example of “Best Practices” within the XML world. We work with the Architecture group and Special Interest Groups (SIG) that are looking at XML as an option. If the hype around XML is wrong then the error will be on timing and impact. The timing is too soon and the impact will be far greater than currently imagined.

My second view comes from the fact that XML is hyped and there are problems in deploying the technology, especially around the scalability. In our experience XML is wonderful for local integration projects. However, it really doesn’t scale very well based on the current technology. This problem will be overcome in time, so get ready.

Wilshire: How about the Dublin Core as a metadata management standard? Where did you find it useful? Does it have a place in mainstream corporate data environments?

Stephens: The value of the Dublin core is in its simplicity and extensibility. The biggest benefit of the Dublin Core is that the standard can be implemented and is an excellent tool to integrate assets. I would much prefer to wrap 150,000 assets with the Dublin Core standard than 10 assets with some extensive telecommunications ontology. Utilizing search engine tools, this standard can provide a level of integration that can’t be duplicated by any other meta-model that I have seen. Keep in mind, the Dublin Core was created to improve resource discovery on the Web. The standard is simple, intuitive, cross–disciplinary, international, open and flexible. All of the elements are optional, repeatable and extensible. All of which make it perfect for integrating metadata assets that are delivered via the web on a corporate Intranet.

Wilshire: And how about your experience with messaging and web services?

Stephens: Let me put a plug in for the DAMA and Wilshire Conference again. This time last year we had not even started looking at message structures and web services. I am sure we would have come around eventually. However, last year’s San Antonio conference provided me with some valuable insight where one company had expanded their meta-model to include message structures. I kicked myself all the way back to Atlanta for not seeing that opportunity.

Web Services is another story. We built our Web Service Repository last year anticipating an enormous growth in the development of web services. The good news is that we had the repository up and running in four weeks. The bad news is that web services have not taken off as we had hoped. When they do, we will be ready.

Wilshire: Todd, you have 6 patents pending in the field of metadata. Can you tell us anything about these or would that put your patents at risk?

Stephens: Before I talk about the patents, let me say this about the concept of innovation. Innovation is not a product of divine intervention. Innovation, like advanced degrees are simply a result of setting goals and a daily attention of thought. Each year, I set a personal goal of applying for 2-4 patents and getting published. We should all set the highest expectations for ourselves; otherwise we will be limited to the expectations of others.

Our third patent application addresses one of the most fundamental problems that you face in an enterprise metadata effort. The issue is that you may not have the field to field mappings that are available when you use an ETL tool. When you offer services across the corporation, you can’t limit yourself to a requirement of ETL architecture. However, this brings up a problem that without the field to field mapping how can you provide impact analysis? One of our patents provides a predictability model for impact analysis based on other factors such as naming standards, system interface definitions, structures, and many other data architecture components. By using the model we can predict impact analysis at the field level even with or without having an ETL product.

Wilshire: And what is Semantic Zooming?

Stephens: There are three basic types of zooming. Geometric zooming allows the user to specify the scale of magnification and increasing or decreasing the magnification of an image by that scale. This allows the user focus on a specific area and information outside of this area is generally discarded. A great example is mapping software like MapQuest or Yahoo. The fisheye zoom is similar to the geometric zoom with the exception that the outside information is not lost from view; this information is merely distorted.

Semantic zooming approaches the process from a different angle. Semantic zooming changes the shape or context in which the information is being presented. An example of this type of technique is the use of a digital clock within an application. In a normal view, the clock may show the hour of the day and date. If the user zooms in then the clock may alter it’s appearance by adding the seconds and minutes. If the user that zooms out, information is discarded with only the date remaining. The actual information did not change, only the presentation method.

What does this have to do with Metadata? The answer is everything. If a user begins to review information at the system level and then drills down, we don’t just add more details about the system. We add interfaces, components and data. If the user drills into the data then we will add the logical, conceptual and physical views of the data. We are actually altering the view in hopes to increase the understanding of the information being presented.

Wilshire: So are you planning to submit yourselves for the Wilshire Meta Data Best Practices Award this year? Did you learn anything in the process of nominating yourselves last year, when you were a finalist?

Stephens: Absolutely, we have been working on this process for 12 months. I was able to have some great conversations in San Antonio with a couple of the other finalists. After talking with them, we identified a couple of weaknesses in our implementation that needed to be corrected in 2002. We have made enormous advancements in the data architecture, usage, content and utility that our implementation provides to BellSouth.

The biggest thing for me as a finalist is the appreciation of being peer reviewed. In the academic world, the biggest accomplishment is to be published in a peer-reviewed journal. The Wilshire award is just that, a peer reviewed award that goes to the organization that has demonstrated “best practices” in the world of metadata. The award generates enormous amount of pride in the organization, verifies our metadata strategy and provides an excellent opportunity to brand the group. We were very fortunate last year to be one of the four finalists. However, behind that fortune is an enormous amount of work, commitment and dedication by the Metadata Services Group. Being one of the four finalists isn’t everything, but wanting to is!

Whether we win, place or show, I for one will be looking to track down and talk to the finalists in Orlando this year. I am a firm believer in the “hire the one that beat you” tactic. These four organizations will have a tremendous amount of knowledge, experience and success in implementation. It simply is a great opportunity to learn from those organizations that believe in Metadata.

Wilshire: OK Todd, well thank you for sharing your insights with us today.

Join us for the
Wilshire Meta-Data Conference
and DAMA International Symposium
May 2-6, 2004 • Century Plaza Hotel • Los Angeles, California USA

The World's Largest Vendor-Neutral Data Management Conference

The 16th annual DAMA International Symposium and 8th annual Wilshire Meta-Data Conference will be held May 2-6, 2004 at the Century Plaza Hotel in Los Angeles, a beautiful venue adjacent to Beverly Hills. Hear 40 case studies outlining strategies of companies that have implemented successful data management projects. There will be more than 120 speakers in all, covering meta data, enterprise architecture, data and process modeling, unstructured data, business rules, data integration, XML, business intelligence, data warehousing, information stewardship, and more. Keynote Speaker Chris Date. Click here for details.