By default, Google Analytics data stored in Google BigQuery table is the source of both transaction and user behavior data.
If you want to attribute the value of products that were bought not online (e.g. brick-and-mortar, call center, etc.), add a transaction data source on your model's page. This may be data exported from your CRM system.
Learn more about importing transaction data to Google BigQuery from this article.
How to add CRM transactions data
1. On the model page, click Add source:
2. Click CRM transactions:
3. Select a Google BigQuery view or table containing data on the transactions you need:
Note:You can connect only those data tables that are located in the same multi-region location as the rest of the source data for this attribution model. For example, if the data already connected to the attribution model are located in the EU, you won't be able to connect data located in the US, and the other Way around. You also can transfer the BigQuery data between the data set in different locations as discribed in this article.
4. Click Save and calculate your model.
Table data schema
Data schema in a Google BigQuery table or view must exactly match this definition.
During the model calculation, data from the source you add replaces data from Google Analytics in the
product_id fields. Only transactions with the completed status are taken into account.
Important:It's critical to have the consistent transaction, user, and product data in both sources.
How does the User ID data correlate between the Google Analytics and CRM sources
Note that if User ID information is not available, you can fill in Client IDs into
user_id field to match data by Client IDs.
|Google Analytics||CRM transactions||How the data correlate|
|userId 1||transaction 1||userId 1||transaction 1||Both user and transaction data match. Actions before the purchase will be assigned to this user.|
|userId 2||transaction 2||userId 3||transaction 2||The transactions match, the users don't. Actions before the purchase will be assigned to the user that was recorded first.|
|userId 4||transaction 3||null||transaction 3||The transactions match, one source has no user data. Actions before the purchase will be assigned to the user that was recorded first. If this will be the 'null' user, only the purchase step will appear in the model's funnel.|
|null||transaction 4||userId 4||transaction 4|
The data is taken from the earlier transaction record among the two sources, GA and CRM.
These correlations work only if the conversion step is a CRM transaction. If the conversion step is a custom event, the conversion and event data will be taken from the custom events table.
How does the transactions data correlate between the Google Analytics and CRM sources
- There is a transaction in Google Analytics, but the CRM transactions data source doesn't contain one
Value of such transactions isn't attributed.
- There is no transaction in Google Analytics, but the CRM transactions data source contains one
- For such transactions, purchase sessions are generated. The sessions have the medium set to 'offline'.
- If a user had online interactions prior to offline purchase, those online interactions get value.
- Value for all funnel interactions that didn't happen online is attributed to the offline purchase session.
- If there were no online interactions prior to offline purchase, all value of an offline transaction is attributed to its session.
- There is a transaction both in Google Analytics and the CRM transaction data source
All information about such transactions and sessions that led to them matches Google Analytics data.