Creating Segments
How to use the Audience Hub visual builder to create and update segments
Last updated
Was this helpful?
How to use the Audience Hub visual builder to create and update segments
Last updated
Was this helpful?
Segments are built on top of datasets, the core models set up for a data warehouse connection. Typically a member of the data team will need to first set up the datasets for your warehouse connection before getting started. Take a look at for details.
To create your first segment, click on Segments 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.
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 dataset, as well as any related datasets 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 datasets you're segmenting, you can also filter based on other data related to this entity.
Related Datasets is the list of other datasets 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 dataset 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.
is null / is blank
All
Column is NULL (or the empty string if given a text column)
is not null / is not blank
All
Column contains any value other than NULL (and not the empty string, if given a text column)
is
All except boolean
Exactly matches the given value (case sensitive)
is not
All except boolean
Does not match the given value
starts with
Text
Starts with the given value
ends with
Text
Ends with the given value
contains
Text
Column contains the given value
doesn't contain
Text
Column doesn't contain the given value
more than
Number
Column is larger than the given value
less than
Number
Column is smaller than the given value
is true
Boolean
Column is true
is false
Boolean
Column is false
more than
Datetime
Column's value is more than ___ days ago
less than
Datetime
Column's value is less than ___ days ago
exactly
Datetime
Column is exactly ___ days ago
after
Datetime
Column is after the given date
on
Datetime
Column is exactly the given date
before
Datetime
Column is before the given date
between
Datetime
Column's value is between ___ days and ____ days ago
contains any of
Array
Array column contains any of the provided values
When the Dataset you're segmenting has a related Events Dataset, those events have even more conditions available, powered by the schema columns defined on the Event Dataset.
Event Name
Any Event
Dataset has any event matching the other conditions
No Event
Dataset has no event matching the other conditions
Event Name is
Dataset has an event with the given name
Event Name is one of
Dataset has an event with any of the given names
No Event Named
Dataset has no event with the given name
No Event Named any of
Dataset has none of the given events
Engagement
At least
Dataset has at least x matching events
At most
Dataset has at most x matching events
Exactly
Dataset has exactly x matching events
Any number
Dataset has any number (> 0) of events
Time Period
Between
Events occurred between ___ and ___ days ago
Prior To
Events occurred before a date ___ days ago
Within Last
Events occurred within the last ___ days
Between dates
Events occurred between two fixed dates
After date
Events occurred after a fixed date in time
Before date
Events occurred before a fixed date in time
Any time
Dataset has match dates regardless of time
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!
At any point, you can also view a sample of the members of a particular segment by pressing the Preview Results button in the top right of the segment view. This will show the first 25 members of your segment as currently defined.
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.
Dimension limits can be placed on a segment to confine the max number of records that share the same value for a specific column. One example is when we want the segment to target no more than 100 people per U.S. state. This can be achieved by setting the dimension limit grouped by the state column.
Setting a dimension limit follows a similar set of steps as segment limits. Enabling the dimension limit is done by clicking the ... menu item in the upper right corner of the segment builder page.
Enter the number of records to limit by
Select the field whose value will be used to limit matching records by. Columns holding enum or non-unique values work best
Select the field that will be used to order by
Select whether you want to order by ascending or descending. This value will dictate the ordering of records within each dimension group
Note: Dimension limits are currently only available for Snowflake and Databricks warehouses.
Census customers often need to manage hundreds of segments. Defining these segments and managing memberships across so many segments often requires defining rules that span across many segments, rather than individual sets of segment rules. Census provides a few different tools to support companies managing large sets of segments.
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.
can be used to specify the other segments that should be considered when defining a segment.
allow defining a "waterfall"-style set of priorities across segments in situations where members may appear in multiple at once. Priorities enforces that a particular contact member will only appear in the highest priority segment they qualify for at any given time.
allow data teams to bulk define segments using SQL and have them automatically imported into Census Audience Hub.
