This use case looks at how to create a classification data stream using a parent-child setup.
RentACarWithUs is a car rental business who wants to classify its Facebook data using a classification that only exists in their internal system. Since this particular classification isn’t part of the data that comes from Facebook, it must be uploaded into Intelligence as a separate data stream. It can then be applied to the Facebook data accordingly.
The finance department has a long list of PO (Purchase Order) numbers in their internal system, which must be added to the campaign data. Create a data stream consisting of the relevant entities: Entity Key - Campaign Key - along with the Classification Dimension - the PO Number.
To map the fields in the classification data stream, use this logic. Since the Campaign Key field in the classification file is identical to the Facebook Ads Campaign Key, it can be mapped to the Intelligence field by the same name. The PO Number is an attribute of Campaign Key, meaning there’s only one PO Number for every Campaign Key. Therefore, it can be mapped as a Campaign Custom Attribute and have its name changed as needed:
| Classification File Dimensions | Intelligence Fields |
|---|---|
| Campaign Key | Campaign Key |
| Campaign Name | Campaign Name |
| PO Number | Campaign Custom Attribute |
The next step is to update the PO Number values (Campaign Custom Attribute) within Facebook, with the PO Number values from the classification data stream. This is done for the same Campaign Keys, as shown in this diagram:
After the data is updated, you can see that relevant PO number is assigned to each campaign within the Facebook data stream:
Now, you must establish a parent-child relationship between the two data streams, so that they’re joined on the basis of matching Entity Keys. This join allows the Facebook data stream to inherit attribute values (Campaign Custom Attribute) from the classification data stream. All PO numbers are automatically updated in the Facebook data stream for matching Campaign Keys.
In this example, set the classification data stream as the Parent and the Facebook data stream as the Child. Any new data sources that are added, other than Facebook, automatically inherit this classification as well.
You can also set up the Facebook data stream as the parent and classification as the child. But, this setup is less scalable and if new data sources are added, each data source needs their own classification data stream.