Attribution Modeler | partitions

The properties in the Attribution Modeler | partitions | partition[n] category specify values that affect how Attribution Modeler evaluates data. There is one set of these properties for each partition.

The enableAttributionModeler property indicates whether the response attribution methods in Attribution Modeler is applied to your reports. Setting this property to False disables the SIRA, first touch, and equal credit attribution methods. Direct and Last touch attribution methods are still enabled to support reports that are a part of Interaction History.

Even though you can change this property value from true to false, it is expected that you set this value only one time in the beginning and never change it. If you want to change this property value after the first run, delete the IH_AMTracking table beforehand to prevent some of the SIRA credits from being left unchanged. Then, rerun scoring in the new enabled or disabled state.

The enableBackgroundCredit property is used to represent all the unknown precursor events of a response with a single "background" offer and compute a proportion of the credit for it. Instead of ignoring the cases where a customer response does not have an earlier contact, the background credit is used to represent all of the previously unknown events of a response. A false contact is created for each customer automatically whenever a customer has at least one response, whether they have any contacts or not.

Even though you can change this property value from true to false, it is expected that you set this property only one time in the beginning and never change it. If you do change the property value between runs, clear the tracking table before changing this property to prevent scoring issues.

This property is enabled only if enableAttributionModeler is set to True.

The Training Period property sets the number of days, looking backward from the current date, for which contacts and responses are pulled from Interaction History to train the Attribution Modeler model during each run.