On March 17, 2016, I held a Real-World Data Governance (#RWDG) webinar with DATAVERSITY that was titled The Data Model as a Data Governance Artifact. The webinar was well attended and the interaction with the participants in the webinar was phenomenal. The Data Modeling crowd is very well educated in the throes of Data Governance. That did not surprise me one bit.
The emphasis of the webinar was on how to make the end result of our data modeling efforts more accessible, understandable and valuable to appropriate audiences throughout the organization when it comes to governing our data. The feedback I received was – that it was not necessarily the model that people thought would be of interest to these audiences. It was what was in the model – the metadata about the definition and structure of the data – that people need to help them to understand and use the data effectively.
The age old question from my early days in the data management industry was – Should Data Modelers show the data model to business and technical people or should they keep it to themselves? Maybe you have a good answer to that question. In my humble opinion, I believe that there are appropriate times to share data models with select people at select times in the data and project life-cycle. It is important to share the contents of the models with business and technical people so that they improve their understanding of the business and technical definition of the data they define, produce and use as part of their everyday job.
In the webinar I laid out an operational model for the roles associated with a Data Governance program. These roles include Executive, Strategic, Tactical, Operational and Support roles. I use this model quite a bit and an image of the model is shown below.
For the balance of this column I will share with you my thoughts about how the metadata in the data model can be shared as the artifact to assist different audiences associated with your Data Governance program.
Sharing the Data Model with the Executive Level
The Executive level of a Data Governance program typically consists of Senior Leadership and may take the form of a Steering Committee that includes individuals at the top of their respective parts of the organization. This level, if it exists associated to Data Governance, often has the responsibility to select the people that will represent their part of the organization at the Strategic level of the Data Governance program. The people they select are often referred to as the Data Governance Council.
The Executives interest in the data models may be limited. This often depends on the size of the organization and Executive involvement in operational issues. Best practice indicates that the Executive level must support, sponsor and understand the activities of the Data Governance program, team and Council for a program to be successful.
It is good for the Executive level to know 1) that the data models are being created and managed, 2) that the models are a part of the overall data management discipline and 3) how the information collected in the models improves the design and understanding of data across the organization.
But the question here is – Should we share data models with Executives? In many cases, the answer is no. Perhaps you share your data models with your Executive team? Perhaps you don’t? Please share (in the comments area below) whether or not you share your data models with people at this level.
Sharing the Data Model with the Strategic Level
As mentioned above, the Strategic level of a Data Governance program often is labeled as the Data Governance Council. The Council typically consists of people that represent the areas that are represented at the Executive level (i.e. there is a person from Finance, a person from HR, IT, Sales/Marketing. Ops …). This group is the top of the pyramid. Issues get escalated to this level when they cannot be resolved at the Tactical level.
The Council’s interest in the Data Models may also be limited. This depends on the Council’s involvement and awareness of the project and program level activities that are taking place in the areas under their guidance. A Council can be used just to resolve issues or they can actively approve artifacts and documents that further the practice of Data Governance within the organization. Some of these artifacts and documents include policies, procedures, standards, glossary data definitions and project specific documentation.
If the Council shows interest in project specific documentation, the data model itself or the information included in the model may be of great interest to them. In this situation, the model can be shared but only with details that match the Council’s interest (conceptual, logical, physical models with all of the denotations included).
So again the question here is, Should we share data models with the Data Governance Council? Perhaps you share your data models with your Council? Perhaps you don’t? Please share again (in the comments area below) whether or not you share your data models with people at this level.
Sharing the Data Model with the Tactical Level
The Tactical level of a Data Governance program is the most active and the most critical to fill appropriately. This level consists of Data Domain Stewards (stewards with accountability associated with a specific domain or subject area of data) and Steward Coordinators (those people in a single part of the organization that engage the stewards through scheduling and communications). The Domain Stewards are often referred to as Subject Matter Experts and they are the go-to people and facilitators when it comes to resolving issues pertaining to their domain of data across the organization. An example of a Domain Steward may be the Registrar for Student data at a University.
Data Domain Stewards are typically very involved in projects and programs involving their specific subject matter of data. Therefore the data model becomes a very important artifact for people at this level. The Coordinators may have a lesser interest in the models and the contents of the models because their governance responsibilities are typically separate from the modeling and project activities.
Data Domain Stewards will most likely be involved in the data modeling activities at your organization. They are a highly knowledgeable, perhaps the most knowledgeable, person in the organization regarding their subject area of data and may depend on the information collected in the data models if their responsibilities include making enterprise decisions associated with data in their domain. Also, it makes perfect sense to include subject matter experts in your data modeling efforts.
Sharing the Data Model with the Operational Level
The Operational level of a Data Governance program consists of… Perhaps every single person in your organization. This depends on the approach that you take to implementing Data Governance. In the Non-Invasive Data Governance approach that I write and speak about a lot, everybody that has a relationship to data – that is, everybody that defines data as part of their job, produces data as part of their job, and/or uses data as part of their job (likely everybody does at least one or more of these activities) is held formally accountable for that relationship with the data. Sometimes the people that manage “everybody else” in the organization can be considered the Operational level of a Data Governance program. Many organizations do not yet recognize that data-related behavior must be influenced at the Operational level in order for a Data Governance program to be successful. This may not be your focus yet, but at some point it may well be.
People at the Operation level are typically more interested in the contents of data models than with the models themselves. This will also depend on your organization. Data models (or the metadata in the data models) can certainly assist in improving the understanding of the data, the structure of the data, the uses of the data and how specific data relates to other data (logically and physically). Data models can be a large part of any data-related effort (read ‘project’) that involves people at the Operational level. Data models can assist Operational people in their day-to-day activities and in their project-related activities.
Operational level people (the Data Stewards) may either have limited interest in the models themselves or they may be actively involved in the development and maintenance of the data models. Therefore it seems to make sense that Data Modelers share data models with people at this level.
Sharing the Data Model with the Support Level
The Support level of a Data Governance program consists of the Data Governance Team (or Leader) and the areas of the organization that support and will be supported by the Data Governance program. The Support level often includes people from Audit, Legal, Information Security and other organizational functions). There is often level of governance (execution and enforcement of authority over the management of data and data-related assets) that already takes place or is generated from these areas.
Sometimes the non-Data Governance Team members of the Support level have a passing interest in the data models and the information included as part of the data models. Team members may be actively involved in facilitating the processes associated with model development so their interest is direct and present constantly).
Supporting areas like those mentioned above will want to know that the data models and associated metadata exists and may ask to see the models at specific points in time. Your Audit team may have a need to make certain that this information is collected while your Privacy area may need to see the details of the models to differentiate between PII and non-PII data. These are examples of why the Support level of your Data Governance program may be interested in seeing the data models. A lot depends on how you structure and who you involve in your Data Governance program.
The webinar that I mentioned at the beginning of this column is available for all to see through DATAVERSITY’s on-demand webinars. Data Modelers are a very active audience and always have something to say or share about their craft. From my experience I see how Data Modelers and data models are under-utilized resources in many organizations. Perhaps if organizations consider new and progressive ways of sharing information in their data models, data models will at some point be considered a dominant artifact to support Data Governance.
What are your thoughts?