Authoritative data source (also referred to as System of Record) is an interesting concept I learned about a while ago and began using immediately. It’s the idea that any piece if data (a person’s name, email address, phone number, etc.) ideally should be stored in only one place, and if it needs to be duplicated, typically for improved performance, one of the locations should be considered the correct, or authoritative, one.
If stored in more than one location, any workflow which changes that information should update it in the authoritative location and then update it in any secondary locations. Alternatively the workflow could update the authoritative data and then a batch process could update any secondary locations.
As the graphic suggests, this concept helps address item 3, accuracy, of data.