Raw facts are, simply said, raw data. The most important attributes of these kinds of data are being unprocessed and not being manipulated in any kind of way.
The Data Vault architecture uses two kinds of rules for raw data: Hard business rules and soft business rules. These two rules affect how the raw data is handled.
Hard business rules do not change the content or any other information of the data.
Soft business rules changes and interprets the data into information.
Because of the Data Vault architecture the raw data is not in the conformed format, which turns the vault into a system of records.
This feature is essential to make the vault auditable. This is reasond by the EDW containing all the raw data in historical format.
The Data Vault Model also needs to be decoupled from the staging area and the information marts. These two can change or even be shut down, but because of the Data Vault architecture all the needed data are available to the EDW.
This makes the EDW auditable and no data get lost over time.
Inmon, Linstedt: Data Architecture: A Primer for the Data Scientist, Elsevier, 2015, S.152ff