Is your Data Vault speaking your language?
Updated: Mar 12, 2021
One of the first lessons I ever learned about data was in the first three days of my data journey. At that moment I did not even know that my professional career was about to change. I was a CAD drafter at the time, and I was placed on a project that was to extract data out from our CAD application. But before we started the project the Data Architect wanted to give us all an intro to data design. To this day the lessons I learned in those few days are still the core foundation to everything I do in data. After that experience I was sold and at that point my data journey began.
Bill Moore (Hi Bill if you are reading this!) was just an independent data professional from coastal Maine but was a true data enthusiast. He wanted others on that project to
be just as passionate about data and to understand what we were about to
venture into on that project. Over those three days we did not talk about the technology or any of the physical aspects of data or databases. We purely discussed the business and how aspects of the business related. Bill identified and shared with us that the most important part of a good data design was an understanding of the business and building that model that accurately represented the business.
The reason I feel what I learned over those 3 days was so valuable is, that over the last 25 years of my data journey. Just about everything technically has changed. I have seen and worked with many different technologies and languages. But what has stood the test of time is the need and importance to have an understanding of that conceptual model of your business. Because regardless of the physical technology, the fact that your business says a Customer can have One or Many Orders and an Order can only have one Customer is a business rule that holds true regardless of the technology you physicalize that data design on.
So, Keith… what does this all have to do with Data Vault 2.0? Well, I think it has everything to do with it.
When starting your data vault initiative, you do not want to just jump in and just start building Hubs, Links & Satellites. You need to take that top-down approach and work with your business SME’s or stewards and define the business ontology and create a business taxonomy.
No, I am not going to turn this into and in-depth dragged-out discussion on what an ontology is vs a taxonomy. But most importantly you need that enterprise-wide perspective of your business from a data perspective. Taking the time to do this exercise will give your data vault the context in a language the business can relate to and understand. Taking the time to do this exercise will affectively help you identify business keys, relationships, descriptive and changing data.
Now you do not need to place your Data Vault implementation on hold until you have done this model for your entire enterprise. But identify your core subject areas or domains and work your way through them in alignment with your business needs and roadmap of what sources of data you will be looking acquire.
One last thing I do want to point out and emphasize is that no where did I say for you to take a look at your source system data model. The whole purpose of why we are building our Data Vault is to empower our business users to be able to drive insight in an effective and efficient manner. That means it needs to be in a language they can relate to and align to their business terms. We are not looking to build a “Source System Data Vault”. You want to be taking a top-down approach to modeling your Data Vault and not a bottom up approach to modeling your Data Vault. We will get to doing that bottom up approach to mapping from you source system to your Data Vault model in a future post.
So remember take the time to define your taxonomy and depending on the size of your organization and if your organization has encountered merges and acquisitions. This could be an interesting adventure. But an exercise that is truly critical for your Data Vault success and if your lucky their might even be some industry standard ones out there to give you that jump start. For more information on Taxonomies and how they work within Data Vault 2.0, I suggest the book by John Giles - Elephant in the Fridge
The key to unlocking the power of data