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

Published in TDAN.com October 2006

Ever experienced this? The phone rings and you hear that a database table needs to be built. They have a pretty good idea of what columns are needed, but would like you to review their ideas and assign standardized names. They’d like to begin testing in a few days, so you need to move this along quickly, please. Then, you open up your favorite data modeling tool, crank out the model and pass it along to the DBA quickly. Even though some of the model doesn’t make complete sense to you, but there isn’t time to quibble over details. Just go with the flow.

Or how about this? An email arrives describing a new project. You get involved early, designing a conceptual model for the project. The design is expressed in the canonical language and has the flexibility of a gymnast. Then the lead goes cold and you don’t hear anything. A while later you hear that the system was built, without any further involvement from the data management team.

These are a couple of the ruts that we data professionals can fall into. We may become database designers, cutting out time for understanding the data to satisfy impending target dates. Or we may spin into theoretical design, losing touch with reality. It’s happened to me: I’ve designed data structures I didn’t fully comprehend. I’ve unveiled data models whose beauty rivals runway models, only to be met by blank, uncomprehending stares. It’s easy to do, and I’m guessing I’m not alone.

How did we get here? Sometimes we fall into these ruts because it’s comfortable and safe. We’re good at what we do and we do it efficiently. Why fix what isn’t broken? Sometimes we are placed into the ruts due to project methodologies or co-worker expectations that dictate what we do and what we don’t do. There is no support for going beyond project deliverables. Why rock the boat?

We need to challenge what we do and adjust our approaches to stay fresh and sharp. We need to force ourselves out of our comfort zone to keep our approach to data management relevant to our customers. We need to convince the world that data management is more than building database tables or drawing fluffy conceptual diagrams.

Breaking out of these ruts means trying new things and seeing what works. Here’s a list of rut-breakers that may help you get out of the rut and into the groove of good data management.

1. If you only have high-level conceptual models, map them to their physical implementations. This will show which concepts are implemented and which ones are not. It will also point out overlaps where the same concept is implemented in multiple platforms across the enterprise.
2. If you only have detailed, physical data models, roll them up to create high-level conceptual models. This will show the semantic content of the databases. It will also point out overlaps where the same concept is implemented on multiple platforms.
3. Acquire, build and retain allies. Make a list of the people in your enterprise who “get” it. Meet with them to understand the value they see in data modeling and build on these strengths. Do the same for those who don’t “get it”: find out what they don’t understand and adjust your approach to address these areas.
4. Rethink data stewardship and governance. With Sarbanes-Oxley, HIPAA and other regulations taking effect, corporate attention is focusing to these efforts. It’s no longer as much about starting a data stewardship program as it is about connecting with the effort that is already underway. Show stakeholders how this work is another tip of the data management iceberg.
5. Go retro and design a data model of your corporation’s hierarchical or network databases. If they’ve survived this long, these data structures are likely supporting core systems that fewer and fewer people really understand. You are a metadata lifeguard: dive in to save this knowledge, transferring it from key employees’ brains into sharable documentation. Capturing metadata on these systems should rise to mission critical status.
6. Look beyond structured data into the unstructured world. Estimates are that 80% of data are stored in an unstructured format, such as policy rates or product features documented in Word documents or HTML pages. Develop a way to catalog this content and maintain that metadata, and you will have built a better mousetrap. (Then write a paper on it for the rest of us!)
7. Become active with DAMA International. Whether it’s through the educational meetings or the informal networking, fresh approaches and new ideas abound whenever DAMA folks get together. Visit www.dama.org to find a chapter near you and plug into this renewable resource!

Incorporating one or two of these ideas will help you get out of your rut. Then, the next time the phone rings or an email arrives, you’ll be able to get involved in time to really add value to the process, and get into the groove of inspired data management.

This column written by John Schley, President, DAMA International.