By: Chris Novak, Technical Consultant
Whether you use Blackbaud CRM or another nonprofit constituent management solution, you may find the need for a data dictionary.
What is a data dictionary?
A data dictionary is a list of key business terms and metrics with specific definitions for your organization. On the surface, this sounds pretty simple - but it carries the ability to reduce confusion and improve data literacy within the organization. It is a valuable tool that a business intelligence/data team can provide to the organization.
Most organizations have at least one term or metric that might be interpreted differently among different departments in fundraising. When someone asks about the funds raised in a quarter, your Annual Giving team and the Major Gift team will likely have varying definitions. For example, does the total value of gifts in this quarter mean the total value of booked gifts, including pledges and gift commitments, or actual cash that came in the door?
Reports may show different numbers for the same metric from the same data source. A data dictionary will clear up this inconsistent and confusing business logic. It serves as a living document that everyone in the organization can reference. From a data/business intelligence perspective, a data dictionary provides more precise reporting requirements and KPIs.
How to get started
For the data/BI team, there are some best practices to go about setting up a data dictionary -
Collect Key Business Terms
Start with a list and go through all the business concepts, dimensions, and measures. An easy way to break this down is to meet with each department and collect their terms.
Define the terms you collected in step 1. To save time, go through existing documentation/requirements and pull any already defined definitions. As you write the description, remember that this is not an internal definition for the BI team but one that everyone in the organization should be able to understand and reference.
After doing the first pass with existing documentation, circle back to each team and get their help to fill in the gaps or refine what is already there; for definitions, you may have gathered from existing documentation, don't take that as gospel; instead, ask for clarity on how this should be defined? Perhaps the existing definition is not ideal or is no longer precisely what they want. This is an excellent opportunity to receive the perfect explanation from the team.
As stated earlier - the Annual Giving team and Major Gift team are likely to be using Total Fundraising as a metric. Go through the working document and identify all of these types of overlaps.
Get Alignment/Come Together
When the exact term or metric is being used by different departments, get together with those departments to review each word. Like the example above with Annual Giving and Major Gifts, it is clear that both of their definitions are unique to them. In this case, each of their terms should get a new name to prevent confusion in future reporting/KPS. An example would be AG Total Fundraising and MG Total Fundraising.
When multiple departments are using the same term, and there is no apparent business reason for why the definitions differ, work to get the teams to agree on a definition that best serves them both. One unit can adopt the meaning of the other, or a whole new definition can be put together that they both use.
Get sign off
Once all the definitions are listed, and the conflicts are ironed out, circle back to the leaders and get buy-in and sign-off.
The data dictionary with all the terms and definitions should be made available to the entire organization and stakeholders.
However, since this is a living document, there are challenges with posting it in a fixed location that may get lost in the shuffle when there are legitimate reasons to update existing definitions or add new ones. Your data dictionary can be published in a Wiki-style tool such as Confluence or WikiHow. Another route would be to load the dictionary into a database table and then surface that table in Tableau® to deliver to the users or perhaps use an intranet like SharePoint.
Deciding where to publish will be different for each organization - it needs to be in a place that is both easily and readily accessible and in an expected location.
As mentioned above, the data dictionary is a living document. May things won't change at all or frequently, but there will always be some business reason or new metrics that may need to be added. It is critical to maintain the dictionary, so the organization has 100% trust in the document. Ultimately, this document's trust level will reflect on the BI/Data team that assembled the record.
Creating a data dictionary is not a small undertaking or a quick win. It requires coordination across all departments and alignment and buy-in from leadership.
That said, building a data dictionary and maintaining it will show significant benefits in the long run. Decisions can be made faster, requirements can be gathered more quickly, and there will be less confusion among reports or metrics. Ultimately, a successful data dictionary will increase the organization's data literacy and open a dialogue and establish a trusting relationship between the business intelligence/data team and the business.