Comment on page
How to use the Audience Hub visual builder to create and update segments
Segments are built on top of entities, the core models set up for a data warehouse connection. Typically a member of the data team will need to first set up the entities for your warehouse connection before getting started. Take a look at Defining Your Data Model for details.
To create your first segment, click on Audience Hub in the left-hand navigation of Census and then click Add a New Segment in the top right.
To start, you'll need to select the data set you're segmenting in the top left. This will determine the type of records your segment contains, the conditions you'll be able to filter on, and what data will eventually be available to sync to your destination tools.
Example of a segment with conditions on the User's email and their events
At any point, you can press Preview Results to get a look at a sample of the data that will be available in your segment. When you're happy with your segment conditions, give you segment a name and press Save.
The basis of any segment is adding new conditions. Clicking Add Condition to select an attribute or related data from the Details List on the right hand side. It contains all the available attributes for the selected entity, as well as any related entities or segments which can be used for additional conditions as well.
On the left side, you'll see the And / Or condition combination control. You can use this to change how conditions are combined. And meaning all the conditions must be met; Or meaning any of the conditions can be met. You can also make combinations of ands and ors by adding Condition Groups which can have their own conditions and combination control.
In addition to creating conditions about attributes on the entity you're segmenting, you can also filter based on other data related to this entity.
- Related Entities is the list of other entities connected to the entity you're segmenting. You can also create conditions on any related entities and your segment will only include records that are related to entities that meet those conditions. Census will take care of building the association between those data sets.
- Inclusions and Exclusions - Allows including or excluding members from a segment if they appear in some other segment of the same entity data. This enables creating sub segments of common shared definitions like Paid Customers, excluding segments that should never be targeted, as well as creation unions and intersections of of multiple other segments.
Census supports a wide variety of operations that can be used to filter segments. The types of operators available are dictated by the types of the data warehouse columns being filtered. Selecting a column with a different type will change the types of operators available.
When the Entity you're segmenting has a related events entity, those events have even more conditions available, powered by the schema columns defined on the event entity.
Though Census's visual segment creator is designed to be used without knowing any SQL, it works by generating SQL underneath the covers and running that against your data. You can actually view this SQL at any point by pressing the View SQL button.
Additionally, if you're trying to create a condition that Census doesn't support, you can still add conditions to your segment by selecting the SQL Condition in the Attribute menu. The SQL you provide will be added as an additional condition alongside any other conditions you've added to your segment. Think of it like formulas in Excel: it's a powerful tool in your toolbelt that can let you express very complex concepts, but it can break the segment so use with care!
A limit can be placed on a segment to confine to a certain number of records. This is accessed when building a segment by clicking the ... menu item in the upper right. The process to add a limit is as follows.
- Enter the number of records to limit by
- Select the field that will be used to order by
- Select whether you want to order by ascending or descending. The top values will be selected so if choosing descending this will be the highest or latest records.
If you want to limit randomly, consider using an ID as the field to order by.
Which string comparison operators are case sensitive and insensitive?
Case sensitive: Census will consider the "is" and "is not" operators to be case sensitive.
Case insensitive: Census will consider the "contains", "does not contain", "starts with", and "ends with" operators to be case insensitive.