Building & Maintaining a Master Data Dictionary

November 29, 2016

Building & Maintaining a Master Data Dictionary

What is a Master Data Dictionary?

A master data dictionary is the authoritative document containing the agreed upon definitions of all the key metrics for an organization. In an ideal world, any metrics being used by more than one person at an organization would be documented in the data dictionary.

Why do You Need One?

A central repository of true and agreed upon definitions of key metrics means your goals will have the same definition across teams, your conversations with investors will be consistent, and your analytical cycle time will be reduced.

At Magento Business Intelligence, we pride ourselves on being data driven. So it's no surprise that a large number of us create our own reports to guide decision making on a daily basis. As a result, we inevitably realized we were sometimes defining the same metric differently across teams. Additionally, as we continued to use data to improve our processes, we sometimes failed to update existing definitions of metrics, leading to inaccurate and inconsistent internal data.

For example, when measuring the customer satisfaction of interactions with our support team, one person was excluding cases where the customer did not provide feedback, while another was considering that a negative response. This led to very different interpretations, numbers, and goals for customer satisfaction among our team. Having a master data dictionary has empowered our team to continue to build their own reports while ensuring consistency and accuracy across the company.

Okay, so I've convinced you that you need a data dictionary, now what?

How Can You Get Started?

With the benefit of hindsight, I can tell you that the key ingredients to creating a successful master data dictionary are:

  • A single person to own the project. Any feedback and updates should go through this one person. This ensures that the design of the dictionary is coherent and uses a single voice.

  • An easy, globally accessible platform. It's important that the definitions are made available to all stakeholders. I recommend keeping the dictionary out of the analytical tool you use because that creates a dependency on that tool, making adapting to a new tool harder. At Magento Business Intelligence, we use a Wiki document. Why? Because we're already using Wiki as a team-wide platform for process documentation and we find it to be the easiest platform to access and modify with version control.

  • A tight feedback loop. You probably won't get everything right the first time. We didn't either. The design and structure of your master data dictionary depends on the people using it, and it's important for their feedback to be considered. Having a single person owning the process helps for this reason; all feedback can be channeled to the same place, and all changes are communicated in the same way by that person.

Who do You Need to Keep in Mind?

It's important to customize the final product to ensure that it's easily accessible by any group within or outside of your organization that is potentially impacted by metric definitions. This may include:

  • Employees. Every employee who uses data is a stakeholder.

  • Managers. When making process changes and strategizing, it's critical that managers are considering the same metrics as everyone else.

  • Executives. It's important for executives to be confident that their top-line metrics have the same definitions at all levels of the organization.

  • Investors. Investors want to know how you define your financial metrics as much as they want to know what the actual numbers are.

This should be enough to get you started! In the next (and final) part of this series, I'll dive deeper into the actual structure of a master data dictionary. While the structure is dependent on the types of metrics, users, and industry of your business, I'll provide some tips and examples from our experience.

About Akash

Akash graduated from the University of Arizona (with its warm and comfortable weather). After college, he worked on a wildly inaccurate poker-playing bot in an attempt to analyze human behavior and then took it upon himself to change the world using data. Akash is an avid follower of international politics and Manchester United.