It would be any Data Manager’s nightmare to run meetings that create pedantic and irrelevant business glossaries or data dictionaries, which eventually gather cyber dust. However, skipping over building and maintaining a good business glossary or data dictionary risks convoluted meanings, confusing communications, and business failures. What should a company do?
Robert S. Seiner (Bob), President and Principal of KIK Consulting & Educational Services, and publisher of The Data Administration Newsletter (TDAN), spoke at the DATAVERSITY® Enterprise Data Governance Online Conference (EDGO) about business glossaries and data dictionaries. To Seiner, formal Data Governance is one of the “four secrets that are necessary build a successful and a sustainable business glossary and data dictionary.”
It Starts with Data Governance and Data Stewardship
Bob Seiner believes creating a good data dictionary or business glossary must start with good definitions of Data Governance and Data Stewardship. Data Governance covers “the execution and enforcement of authority over the management of data and data related resources.” When Seiner spoke of data-related resources, he meant the metadata to be contained in the business glossary and/or data dictionary “or even the DBM’s catalog.” According to Seiner, Data Stewardship means “the formalization of accountability over the management of data, and the data related resources.” Data Stewardship encompasses everybody, as “people have a relationship to the data, just as they have a relationship to the metadata.” As a result, accountability of these people needs to be formalized to capture and record the best possible metadata.
Formal accountability comes in many flavors, but Seiner prefers the Non-Invasive Approach™. This means “applying governance to existing processes rather than redefining all the processes.” But in any Data Governance style, he emphasized, “organizations want to execute and enforce authority. That comes down to how companies implement Data Governance and apply it through Data Stewards.”
Businesses Own Glossaries and IT Owns Data Dictionaries
In addition to grasping Data Governance and Data Stewardship, managers need to understand the major differences, highlighted by Seiner, between a business glossary and a data dictionary. He quoted Nicola Askham:
“The business creates, maintains, and owns the business glossaries. Information Technology (IT) or the people that own the systems take on responsibility for data dictionaries. So, the differences between the two has to do with who’s taking care of it, and who’s maintaining it, and who the Stewards are of that metadata.”
Expanding further, he cited Lowell Fryman:
“The prime directive of the business glossary is to eliminate semantic confusion across the enterprise. This means each business term within the business glossary needs a unique name, a single definition, a single value set, a single set of business rules, an authoritative source. This requires defining terminology used by the business. There may be multiple usages, but often those usages don’t conflict with the single definition in the business glossary.”
In contrast: A data dictionary is created for specific purpose (e.g. to document a specific set of data within an application, Data Warehouse or Data Lake). While business terminology focused, said Seiner, “data dictionaries link these concepts specifically to the metadata about the data or the information in the catalog.” Collecting or connecting business glossary terminology to data dictionary information may happen, and “a data dictionary typically has a business definition in it. But IT has the responsibility to put it there.”
Seiner concluded from his analysis that business must be involved in glossary and dictionary development (the second of Bob’s secrets). Given Seiner’s background on dictionaries versus glossaries, he moved into his four recommendations.
Four Secrets of Successful Business Glossaries Data Dictionaries
1. Formally Govern the Contents
“Have a plan that can be demonstrated and include specific details, such as who will collect the information, what is the process and level of commitment within the organization. Evaluate how well the contents were formally governed,” said Seiner.
Seiner’s corollary to this is: Be sure to pick data that is most meaningful to the organization, because there tends to be so much data it would be “unthinkable to record metadata for every single data element in the organization.” He called this kind of data Common or Critical Data Elements (CDE). He explained:
“They are specific Data Assets. Often business directs the work in collecting and deciding which CDE to use. CDE may come from multiple data dictionaries, multiple places—perhaps even data reported on the company’s dashboards.”
Seiner emphasized knowing data that adds “specific value to the organization through
collecting information about business terms and then storing them physically.” Identifying CDE and following a formal governance plan may not always be smooth. He counsels businesses to “tell when there is a lack of resources, when the process is breaking, and what is the needed time frame to document all of this information,” and then address these to get Data Governance planning back on track.
2. Involve the Business in the Definition, Production, and Use of Tool
Seiner stressed that a successful data glossary or data dictionary requires “getting the right people involved at the right time, in the right way, using the right data to make the right decisions,” in other words, the “Data Governance Bill of Rights.” He believes “enlisting the right people as Data Stewards in the organization” is paramount.
The right people depends on who spearheads the glossary or dictionary development. The ultimate sponsor needs to be known and will heavily determine business’ role. This person in may be I.T., Chief Data Officer, or people in business itself. Once this is known, “the business can understand the role they will play collecting information for their dictionaries and/or glossaries.”
From there, Seiner said, “this well-thought out plan about the roles needs to gain approval from a strategic level of Data Governance.” Buy-in comes from providing details about how to collect information for the dictionary and/or glossary and may include tools like flow charts and RACI (Responsibility Assignment Matrix) charts. Getting the okay for business’ role is essential.
3. Apply Structure and Guidance
As companies start down this path, Seiner suggests they look at how this information will be collected immediately. He noted: “While the market has made lots of tools available, companies considering them need to demonstrate their value. This means focusing on immediate pooling and ensuring the information is collected consistently.” A process for defining, reviewing, and providing feedback on information considered as Critical Data Elements needs to be in place.
Change management workflow needs to be available for dictionaries and glossaries. That way Data Stewards and people on the ground floor of “the organization can participate, ensuring and providing adequate definitions, and addressing the questions raised by the business community about the data.”
4. Incrementally Build and Manage Data Dictionary
Just as a construction company would not build a house in one day, Seiner believes organizations create data dictionaries and business glossaries over time. He stated that the success of this incremental approach comes from “discussion around the reasonable number of elements the we include in our dictionary, e.g. an adequate time frame to put in place. Furthermore, we will want to engage Data Stewards in this approach.”
Building a successful business glossary and data dictionary requires sticking to his four principles: formally governing contents, involving the business in the definition, production and use of tool, applying structure and guidance, and incrementally building and managing data dictionaries. Seiner concluded by sharing a high-level view and a conceptual framework in putting together business glossaries and dictionaries that work.
Check out Enterprise Data Governance Online at http://datagovernanceonline.com/
Here is the video of the Enterprise Data Governance Online Presentation:
Image used under license from Shutterstock.com