By Jason Lim
Published on 2021年4月26日
In your organization, are you ever confused by different definitions of business terms? Do you ever wonder why the number of customers differs between two reports?
I’m going to assume the answer is “yes.” And you are not alone.
As an organization evolves, it’s natural for the language to evolve, too. That’s why it’s critical that important terms be defined, documented, and made visible to everyone. This aligns understanding, so that, for example, a term like “customer” means the same thing across all departments.
This is where a data dictionary and business glossary become useful for getting both your business and IT teams on the same page.
A data dictionary defines and describes technical data terms. Data terms could be database schemas, tables, or columns. It may include information about the data type, size, default values, constraints, relationships to other data, and the meaning or purpose of a given asset.
Data dictionaries are designed for more technical audiences, like IT or data scientists.
These folks will reference the data dictionary to understand data elements, which allows them to manage, move, merge, and analyze data with clarity.
For complex projects, like data wrangling, modeling, or database design, a data dictionary is a helpful resource. This is especially true for new hires. For these tasks, they may look to the data dictionary to ensure use of the right assets.
A business glossary defines and describes business terms and organizational nomenclature.
If you’re thinking “business term definitions” are straightforward, think again. For businesses operating across many departments, regions, and contexts, a single term can have multiple meanings. This can become confusing and result in costly mistakes.
For example, to the sales department, the term “revenue” may include gift cards, but not to the finance department. “Customers” and “users” might sound the same, but they are accounted for differently depending on who you ask. The term “beta feature” may be clear internally, but not to the Customer Success team, who communicate with customers.
Simply put, having a clear definition and understanding of business terms is extremely useful for any enterprise.
A business glossary is useful for the business audience, or people working in functional departments, such as finance, marketing, or sales. When new people join an organization, they can look to the business glossary to learn the business language.
A business glossary helps an organization agree and align on internal definitions. How often have you faced a problem where you think a business term means one thing, but another team believes it means something else? If an executive is presented with two different reports describing the same term, without context on why they differ, how will they react? In all likelihood, they will distrust both reports.
Another benefit of the business glossary is self-service. Users can help themselves without asking around to find an answer. This in itself promotes efficiency and productivity for everyone.
There are many differences between a data dictionary and a business glossary. Below, I’ve listed the main ones:
A data dictionary focuses on physical data assets while a business glossary focuses on business concepts.
The key artifact of a data dictionary is a list of datasets/tables and fields/columns while a business glossary provides a list of business terms and their definitions.
The goal of a data dictionary is to understand data assets and databases while a business glossary’s goal is to define common vocabulary and understand basic concepts.
A data dictionary is owned by IT while business glossaries are owned by the business.
A data dictionary scope is one per data source, while a business glossary’s scope is one per organization.
A data dictionary is applied through data modeling, database design, and documentation of data sources. While a business glossary is applied by data governance and requirement analysis.
A good example of a data dictionary is “USER_ID” (unique 7-digit identifying number of a user). A good example of a business glossary is by “Active Days” (the number of days which a visit to the app is recorded).
I also broke this down in an outline below:
We’ve established the value of both a data dictionary and business glossary. Now the question becomes, how to create them?
On the surface, both sound easy and straightforward to develop. But there are indeed inherent challenges.
For a data dictionary, the volume of new data is constantly increasing. This makes it difficult to keep up with all the new data elements that need to be defined. To keep pace, IT or database administrators may automate data dictionary maintenance. Automation helps IT build and preserve the integrity of the data dictionary.
For a business glossary, it’s important to assign ownership. Definitions may differ depending on context or team. You must ask:
Who is responsible for defining the terms?
Are definitions set by one person or by group consensus?
Is there an approval process?
These challenges require both a robust process and tool to manage a business glossary.
Alation is a machine learning data catalog. Once connected to data sources in the environment, Alation automatically indexes data and populates catalog pages by source. Let’s take a look at how the Alation Data Dictionary and Business Glossary produce and simplify vocabulary complexities:
In the example below, a data dictionary of columns in the “metrics” table has been generated. In Alation, this is what a data dictionary looks like:
This table shows the technical column name, a business title name, the data type, and popularity. Users can click on the blue column links to get more information and context about the columns.
Note that the business “Title” names are auto-titled. Machine learning translates the technical column names into natural human language. For example, “ts_created” has been identified as “Timestamp Created” by the machine and confirmed by a person. This is indicated by the green robot head, AKA Allie.
“Popularity” is an Alation-specific measure of how much the column has been searched and queried by the users. This is calculated automatically through Alation’s Behavioral Analysis Engine (BAE). It does so by accounting for how frequently and recently the columns have been used.
Having insight into the popularity of data is useful. This is because it lets people know what data is most worthy of their attention. It also offers insight into the data that is most trusted, a good indicator of quality.
The business glossary in Alation is more than a place to look up terms and definitions, or assign owners. There are some unique features that make Alation’s business glossary easier to manage and scale. Each glossary is made up of a collection of terms. For example, the “Financial KPI Metrics” glossary may include terms for “EBIT”, “CACC”, and ‘“Debt-to-equity” ratio. The Glossary columns can be customized for different fields, such as”‘Author”, “Description”, or “Status.”
To keep the glossary layout consistent, templates can be applied so that all terms contain the same useful information.
By leveraging Alation’s wiki-like articles, users can @mention and link another data asset, person, or group in Alation. This makes it easy to connect related and relevant information for users to follow.
Typically the process of adding a new business glossary term can seem overly complicated. A group of people may need to come together to debate, agree, and assign an owner to a term.
By contrast, in Alation, a lightweight workflow mechanism called “Agile Approval” assigns reviewers to a term. In other words, anyone can add a term for speed and transparency. However, for that term to be validated, the designated approver must approve the new term, at which point a green “Approved” banner will indicate it is certified.
When creating a business glossary, the sheer volume of terms can present a challenge. Many terms risk being forgotten or missed. That’s why a mechanism to automatically identify new terms that are commonly used within your business is essential.
Alation’s Auto-Suggested Terms does just that. It finds popular terms and their associated technical abbreviations to show the related data objects where they appear, along with popularity. The new suggested terms can then either be added as a new glossary term or to an existing article.
In essence, a data dictionary is for data terms and a business glossary is for business terms. They both have value in aligning technical teams and business teams around a shared understanding. Such alignment translates into setting the right goals, calculating metrics in the right way, and basing it on the right data.
So now that you understand a data dictionary vs. a business glossary, and how they differ, it’s time to implement one into your business.
Join a weekly live demo and see Alation’s business glossary and data dictionary in action.
A data dictionary defines and describes technical data terms.
A business glossary defines and describes business terms and organizational nomenclature.
In data dictionaries, IT typically owns definitions, while business glossaries are owned by business units. Each may have designated roles or departments responsible for maintaining and updating definitions.