Right after you create a model, model “Settings” tab is opened:
Transactional window is a period of time out of which transactions are analyzed. Value is attributed only to sessions that lead to these transactions.
For example, if you need to calculate model for the month of March, set transactional window as 31 days before 2016-03-31 . Leave date field empty to calculate model for the last N days you set.
Conversion window is a period of time before each transaction. Value is attributed only to sessions of this period.
By default this parameter is set to 30 days. But time to purchase can be different for your project. Before changing the parameter we recommend take a look at “Time Lag” report in Google Analytics. It shows how many days pass from the first visitor's interaction to a purchase.
A few examples of how these parameters influence what sessions are get attributed value:
Field “Index of custom dimension that contains User ID” is required if Google BigQuery Export for Google Analytics is a source of your attribution model.
To check that index please open Administrator menu in your Google Analytics account and choose Custom Definitions → Custom Dimensions.
Attribute value of product or transaction
The option defines how resultant data is grouped — by products or by transactions. If you choose to attribute value of products, you'll be able to find out conversion paths for each product with product_id field in resultant table.
If data about products is collected in your project, it's better to select “Product” here. This will give you more accurate results when you analyze results. For example, if you need to consider margin that is set for each product or product category.
Assign value not to all the channels
There is an option to assign value of specific traffic channels to the previous channel in the conversion path. This option is described in more details in the article “Set channels that should not get value assigned”.
After you set up parameters of your model, you can proceed to setup of funnel steps.