You can create custom tags by defining filter conditions as tag values. These conditions can be derived from tag filtering, Recency, Frequency, Monetary (RFM) model filtering, Awareness, Interest, Purchase, and Loyalty (AIPL) model filtering, behavior filtering, or cross filtering. This process generates a custom tag dataset. Custom tag datasets can be used for perspective analytics, audience filtering, and pushes, similar to regular user tag datasets. You can also manage custom tags in the same way you manage regular tags.
A custom tag can have multiple tag values. Each value represents a set of filter conditions. These conditions can be derived from different filter types and datasets. However, all datasets must use the same user identity type.
The filtering process is similar to audience filtering. The user identities from the filtered audience and the custom tag are combined to create the custom tag dataset. The user identity field in this dataset is always named `user_unique_id` and can be used for pushes.
Within a single data source, all custom tags that share the same user identity type and are created by the same user are saved in one custom tag dataset. Each tag is stored as a separate field in the same data table.
Numeric and time-based tags cannot be used directly for perspective analytics. Custom tags provide a solution by allowing you to analyze these types of tags. You can define a range of continuous values, such as a price range, age range, or time period, as a single tag value for a custom tag. Then, you can use the custom tag to perform perspective analytics on that value range.
For example, you can create a custom tag named `Age Range` and define values for this tag, such as `1-10`, `11-20`, and `21-30`. You can then perform perspective analytics on the `Age Range` tag.
Procedure
Go to Workspace > Configuration Management > Data Center > Dataset, or Workspace > User Insights > Asset Management > Tag Management. In the upper-right corner, click Custom Tags to open the page for creating a custom tag.
Set the tag information.
Field Name: Enter a field name for the custom tag. The prefix `user_tag_` is fixed.
Tag Name: Enter a name for the custom tag.
Tag Category: Select a tag category from the drop-down list. To edit tag categories, click Tag Category Management. For more information, see the tag category management section in Manage user tag datasets.
Tag Description: Enter a description for the tag.
Configure tag values.
Set the Tag Value and Tag Value Remark. The tag value should reflect the filter conditions, such as `Age 1-10 in XX dataset`.
Select Tag Filtering, Model Filtering, or Behavior Filtering. Then, select a dataset and set the filter conditions. Cross filtering is also supported. For more information about filtering methods, see Audience filter document.
After you complete the configuration, click New Tag Value. This action saves the current filter conditions as a tag value and adds a new, empty tag value for you to configure.
To remove a tag value, click the
icon next to it and confirm the removal.NoteYou can add up to 50 tag values.
New tag values appear at the top of the list by default. If you have many tag values, click the
icon at the end of the row to view the hidden values.
After you configure all tag values, click Done to save the custom tag.
If no custom tag dataset with the same identity type exists in the current data source, a pop-up window appears. In this window, enter a dataset name. The name prefix is fixed to `Custom Tag Dataset_User Identity Type_`. Then, select a location to save the dataset and click OK.
The custom tag dataset is saved as a data table. The table is named `label_User Identity Type_Ordinal Number` and contains the custom tag field and the user identity field.
If a custom tag dataset with the same identity type already exists in the current data source, the new custom tag is saved to that dataset. A pop-up window shows the name of the existing dataset. Click OK.
You can manage custom tag datasets from the dataset list. For more information, see Manage custom tag datasets.