We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.


The Need for the Universal Translator: The Metadata Repository

Originally published August 25, 2005

When I was privileged to write Data Warehouse Architecture: The Great Debate between Bill Inmon and Ralph Kimball, I had to do a lot of research on both of their philosophies and methods.

One of the things that stayed with me was that, if you follow either of their methods to the letter, you will have a superior result.

One of the requirements for Ralph Kimball’s data marts has troubled me, however. This is his requirement for conformed dimensions. Conformed dimensions usually require that everyone who will use the data will refer to the data by the same name. This actually leads to the concept of a metadata repository to effectively utilize the data warehouse.

This article will relate my experiences as general counsel in a Fortune 100, and several larger multinational corporations, dealing with corporate data systems and their impact on litigation.

In the first instance, unless the corporation is just starting its data warehouse effort, or unless it has gone through some sort of overall corporate trauma, my experience has been that the capturing and storing of the data that the business needs is done by each department.

One would think that this would be the ideal for the data mart approach, except, it would suggest that consistent naming conventions must be used.

The corporations I worked for were heavy-manufacturing and most specifically water and wastewater treatment manufacturers. When data systems were set up, it was mostly department by department. Each manager had, or wanted, his own system.

As an example, if we sold a clarifier; which is a large tank with complex internal machinery to initially start the water purification process:

  • Accounting had a six number and four letter tracking code as an order number.
  • Sales and Marketing used customer location “LA County.”
  • Manufacturing used a numerical code that encompassed the material components.
  • Customer Service used the customer name “Hyperion.”
  • Engineering used the name of the product with a date “Clarifier _______2005”.

Compound this naming challenge with the fact that we used independent distribution and sales representatives who also used different naming conventions in their own business such as sales person, county, vendor, order number etc. One could argue that all of these entities should be required to use the same naming convention. But, we all know that if marketing is using its budget for creating its system, it doesn’t want to subsidize accounting, manufacturing etc. (and vice versa). Also, with regard to the independent entities, our sales representatives and distributors, the corporation had no control over what they chose to use.

When you have one event, the sale of a clarifier, that is classified under so many different names, as long as everyone is used to the process, there is no problem and business proceeds as usual. However, as I said, I managed everything related to legal matters, and almost every one of my requirements was an exception to normal business practices.

Actual Litigation Experience

When my company was served with a summons and complaint, alleging product malfunction or personal injury, the key nomenclature in the summons and complaint was the name of the plaintiff. If it was an individual because of a complaint for personal injury it would never coincide with any of the data names the company used for the product. If it was for malfunction or improper performance, it could be named by any one of our departments' nomenclature.

In a legal situation such as this, because of very stringent time deadlines to answer, or otherwise plead, there are many critical assessments and communications to be made within the corporation as soon as the summons and complaint is served. These are required to preserve evidence of the matter and include:

  • All documents (including electronically stored communications) must be preserved. This means that everyone, inside and outside of the corporation, must preserve the documents related to this action. It falls upon counsel, either inside or retained outside counsel, to send out the notices to preserve these documents. Obviously, as inside counsel, I knew the various departments’ naming conventions, but outside counsel would probably have to scramble to secure this important piece of the litigation. This is where it becomes critical to develop a “universal translator” to marshal the documentation, because if inside counsel leaves, so does this repository.
  • Someone within the corporation will need to be designated as a rule 34 (b) witness. Unfortunately because of the intensity of the knowledge the witness must have, it will most likely be the CEO, CFO or Chief Engineer. This person will need to acquire a familiarity with all of the documents in the case. His or her role is to certify that the documents that have been submitted are genuine and are a full response to the other side’s demand. When the information concerning each product or project is kept under such disparate names as cited above, then the rule 34(b) witness may feel uneasy about the scope of his testimony. It would be very easy to have missed something that can be later raised by finding it in the Plaintiff’s documents. From counsel’s standpoint, there is more to managing this matter than just identifying, copying and reviewing the documentation.
  • Mounting an effective defense—if every department classifies or names the product or project something else, it is very difficult to be sure that an important piece of the defense has not been missed. Also, it may be difficult to assess the extent of involvement for the various departments, if everyone is not focused on this one particular sale.
  • Settlement possibilities—almost all litigation is settled, not tried. If there is no common nomenclature within the corporation, the data will need to be evaluated by each department, on each product or project to be able to determine whether a settlement in any particular case is worthwhile and will not set a negative precedent for the future.
  • As a preventive measure for other litigation — just as multiple classification of the same data makes settlement assessment difficult, it makes it difficult to use the procedure as a learning experience. Being sued is an incredibly expensive situation, both in legal fees and the time management and the other employees are diverted from their jobs. If the corporation doesn’t learn anything from such an experience, the loss is compounded and likely to be repeated.

Is there a possible metadata repository or “universal translator” that every department would use? Would one department volunteer to pay for it? Probably not, because this really needs to be an enterprise-driven initiative. What does the rule 34 (b) deponent do when he finds that because we haven’t correctly specified the documents that a major part of them have been destroyed? An entire company can be lost through “shredding.” We remember Arthur Andersen, don’t we? These are problems that every corporation must address. The vendor community can help businesses develop the metadata repositories so critically needed.

  • Katherine Drewek

    Katherine (1950-2010) had more than 30 years of experience in the editorial and corporate law environment. She was responsible for the content review, editing and formatting of international newsletters focusing on business intelligence and data warehousing. She was a frequent lecturer and panelist for the American Bar Association, the National Association of Credit Managers and the Corporate Practice Institute. She had been a mentor for the Women's Leadership Conference and served as Managing Editor for BeyeNetwork.

Recent articles by Katherine Drewek

 

Comments

Want to post a comment? Login or become a member today!

Be the first to comment!