We often hear the word “metadata,” especially when talking about data migration, system copy and synchronization, stressing the importance of “metadata” alignment.
The short definition of “metadata” is “data about data.”
But I’ve realized that the meaning of metadata is very broad and depends on the context and who you talk to. This creates a lot of ambiguity and misunderstandings when starting initiatives in which data is key, such as when splitting a Salesforce instance.
The first meaning of metadata is the data complementing the data.
Let’s take the example of a photo you take with your phone. The “metadata” of the photo is the date it was taken, the camera settings, the location, etc. This complements the photo, allowing you to better categorize it.
Picture taken on Jun 23, 2022, 2000x3000 pixels, ISO = 400
In a similar way, the “metadata” of a contact in a CRM like Salesforce is the date the contact was created, and the name of the user who created it.
Account “STW” created on May 2, 2022 by [email protected]
But if you take a database approach, this metadata is… data. Just extending the data model with extra context attributes. This is just more data complementing the data model.
The metadata can also mean the “data model,” describing how the data is structured: what are the various attributes, their type (text, date, boolean, picklist…) the controls on them (mandatory, list of possible values, link with another record…).
This can be done at the functional level or at the technical level (database definition).
Name, text, mandatory
Address, text, mandatory
Web site, url, optional
Created on (complementing data)
Created by (complementing data)
This data model is a foundational element of data governance, coupled with a definition of what is an account (which is far from being trivial and may slightly vary between one company and another…)
As Bob Dylan would sing “Man gave names to all the animals, In the beginning, in the beginning”
The third meaning of metadata is analytics about your data.
As soon as you have data, you can start establishing statistics on this data: how many accounts have been created in the last month, who is the user creating the most opportunities, etc.
Which is in turn data… That can be analyzed and used to discover new insights.
In the Salesforce world, the word “metadata” is used to describe what is done to configure the Salesforce Software-as-a-Service and is a collection of:
When discussing data governance, the meaning of “metadata” is most of the time the “data model.” It is necessary to have a company-wide functional data model to enable consolidation and collaboration. And you must have technical data models (such as the Salesforce one) as close as possible to the functional one to ensure that data can be easily exchanged between applications.
This aspect is even more important when there are multiple implementations of the same platform in a company, such as Salesforce or the SAP ERP and the need to share data across them.
In conclusion, let’s try to be more specific when we talk about metadata so we are all on the same page!