Filter builder
The Filter builder lets you build simple or complex filters that identify person attributes and actions across visits and events. It provides a canvas to drag and drop metric dimensions, events, or other filters in order to filter persons based on hierarchy logic, rules, and operators.
For information about how to create quick filters that apply only to the project where they are created, see Quick filters.
Access the Filter builder
You can access the Filter builder in any of the following ways:
- Top navigation: Click Customer Journey Analytics > Components > Filters.
- Analysis Workspace: With a project open in Analysis Workspace, select + Components > Create filter.
- Report Builder: Work with Filters in Report Builder.
Builder criteria overview section_F61C4268A5974C788629399ADE1E6E7C
You can add rule definitions and containers to define your filters. (For information about accessing the Filter builder, see Access the Filter builder.)
Available for dimensions only, these models determine what values in a dimension to filter for. Dimension models are particularly useful in sequential filters.
- Repeating (default): Includes instances and persisted values for the dimension.
- Instance: Includes instances for the dimension.
- Non-repeating instance: Includes unique instances (non-repeating) for the dimension. This is the model applied in Flow when repeat instances are excluded.
For an example, see the “Attribution models” section below.
Lets you preview the key metrics to see whether you have a valid filter and how broad the filter is. Represents the breakdown of the dataset you can expect to see if you apply this filter. Shows 3 concentric circles and a list to show the number and percentage of matches for People, Sessions, and Reports Run for a filter run against a dataset.
This chart is updated immediately after you create or make changes to your filter definition.
Build a filter build-filters
-
Simply drag a Dimension, Filter, or Metric Event from the left pane to the Definitions field.
-
Set the operator from the drop-down menu.
-
Enter or select a value for the item selected.
-
Add additional containers if needed, using And, Or, or Then rules.
-
After placing the containers and setting the rules, see the results of the filter in the validation chart at the top right. The validator indicates the percentage and absolute number of page views, visits, and unique persons that match the filter you created.
-
Under Tags, tag the container by selecting an existing tag or creating a new one.
-
Click Save to save the filter.
You are taken to the Filters manager, where you can tag, share, and manage your filter in multiple ways.
Add containers containers
You can build a framework of containers and then place logic rules and operators between.
-
Click Options > Add container.
A new Event container opens without an Event (Page View) identified.
-
Change the container type as needed.
-
Drag a Dimension, Filter, or Event from the left pane to the container.
-
Continue to add new containers from the top-level Options > Add container button at the top of the definition, or add containers from within a container to nest logic.
OR
Select one or more rules and then click Options > Add container from selection. This turns your selection into a separate container.
Use date ranges date-ranges
You can build filters that contain rolling date ranges in order to answer questions about ongoing campaigns or events.
For example, you can easily build a filter that includes “everyone who has made a purchase over the past 60 days”.
You create a Session container and within it, add the Last 60 days time range and the metric Orders is greater than or equal to 1, with an AND operator.
Here is a video on using rolling date ranges in filters:
Stack filters stack
Stacking filters works by combining the criteria in each filter using an ‘and’ operator, and then applying the combined criteria. This can be done in a Workspace project directly or in Filter builder.
For example, stacking a “mobile phone users” filter and a “US geography” filter would return data only for mobile phone users in the US.
Think of these filters as building blocks or modules that you can include in a filter library, for users to use as they see fit. That way, you can dramatically reduce the number of filters needed. For example, assume you have 40 filters:
- 20 for mobile phone users in different countries (US_mobile, Germany_mobile, France_mobile, Brazil_mobile, etc.)
- 20 for tablet users in different countries (US_tablet, Germany_tablet, France_tablet, Brazil_tablet, etc.)
By using filter stacking, you can reduce your filter count to 22 and stack them as needed. You would need to create these filters:
- one filter for mobile users
- one filter for tablet users
- 20 filters for the different geographies
-
Go to the Filter builder.
-
Provide a title and description for the filter.
-
Click Show filters to bring up the list of filters in the left navigation.
-
Drag the filters you want to stack to the filter definition canvas.
-
Select Save.
Attribution models attribution
Example: Event filter where eVar1 = A