Segmentation
The segmentation capability enables you to group your customers based on demographic, transactional, or behavioral attributes. You can use segmentation to target promotional campaigns, sales activities, and customer support actions to achieve your business goals.
You can define complex filters around the Customer Profile table and its graph of related tables. Each segment, after processing, displays a set of customer table records that you can export and act on.
There are two types of segments:
Static: A segment that is processed only once---either upon the creation or update of any of its filters. These segments are especially useful when properties don't change over time or when they're only used once. Example use case: Customers who attended an expo event.
Dynamic: A segment that is processed according to a recurring schedule. These segments are especially useful when customers' attributes change over time. For example, customers who bought products worth more than $500 in the last three months. Dynamic segments refresh every 12 hours.
Segments are created from the Segments page, where you can view all your saved segments and perform certain actions. Any Dynamic segments appear on the left. Static segments appear to the right. Each segment is represented by a tile that includes the segment's name, description, last date of data refresh, and historical trend (if it exists).
Initially, when you create a new segment, you need to define the following parameters:
Type (Required) - Specifies whether the segment is a Dynamic or Static segment.
Name (Required) - The name that appears for the segment on the Segments page.
Names must begin with a letter and can only contain numbers and letters. No spaces or special characters are allowed.
Display name (Optional) - The display name is an easier-to-read name of the segment to display throughout the application.
The display name can include letters, numbers, spaces, and special characters.
Define a group
When you create segments, your first task is to define how the data in your segment is grouped. Each group contains a filter that defines which records to include in the group such as all customers whose average web purchase value is more than USD 139.00. A filter expression contains four parts:
Table - Defines the table that includes the specific attribute that you want to segment by. This element might be your customer profile table, a measure that was created, or if you ingested a table.
Attribute - Defines the field that you want to use as the segment.
Operator - Defines the comparison operator to use against a value.
Value - Defines the value that the operator is using for its comparison.
Each filter that you define must end with the Customer Profile table. The reason is because the Customer Profile table is where the records are pulled from. You can add as many related tables that you need to refine the results. But your last filter criteria must be on the Customer Profile table. You can also use the All Records operator in case you don't need to slice your data by any of the Customer Profile table fields.
The following two logical operators can be used when you want to include related data:
AND - Both conditions must be met as part of the segmentation process. This option is most useful when you define conditions across different tables.
OR - Under this option, neither of the conditions needs to be met as part of the segmentation process. This option is most useful when you define multiple conditions for the same table.
Currently, it's possible to nest an OR operator under an AND operator but not the other way around.
Important
You need to define all relationships prior to using them in a segment.
Occasionally, you might want to add more groups to help further refine how the customers are segmented. Each group that you define has its own resulting records.
For example, consider a scenario where you defined the following two groups for a segment:
Group 1 - Includes customers with average web purchases that are greater than USD 139.00 (50 records).
Group 2 - Includes customers who live in New York and have a Lifetime Spend that is greater than USD 3000.00 (40 records).
As you add more groups, you can define how the common data between the two groups is maintained. Three options that you can choose from are:
Union - Unites the two groups together as one, which creates a larger group in total.
In the preceding example, Groups 1 and 2 would be combined and the segment would have 90 total records.
Intersect - Overlaps the two groups. Only data that is common to both groups is maintained in the unified group.
In the previous example, the 50 records from Group 1 would be filtered down. So only the records that meet both Group 1 and Group 2 conditions are included. Only New York customers whose average web purchase is greater than USD 139.00 and Lifetime Spend is more than USD 3000.00 would be included.
Except - Combines the two groups. Only data that is not common to both groups is maintained.
In the preceding example, any records that match both Group 1 and Group 2 criteria would be removed. Only non-New York customers whose average web purchase is less than USD 139.00 and whose Lifetime Spend is less than USD 3000.00 would be displayed.
When you save your segment, it's tested to ensure it's valid. If requirements aren't met, it saves as a draft. Segments in draft mode are inactive segments that aren't available for use. Segments can be manually deactivated if you want to make edits or simply don't want them available to run.
If a segment is in draft mode because it contains invalid arguments, you can't run or activate the segment until you correct the invalid arguments and it's considered a valid segment.
Manage segments
Each segment that is defined has different options that can be performed against it such as deactivating it, if needed, or making changes to the list. You can access the available options by selecting the ellipsis on the segment tile. You can select from the following options:
Edit the segment
View the segment's members
Export the segment to either a CSV file or to a Dynamics 365 Sales location
Change a Dynamic segment to inactive or active (depending on its baseline state)
Run a Static segment
Delete the segment
View processing history and segment members
Because Dynamic segment members are going to change, you might want to keep track of how the list's members change over time. Each segment can be opened to view the Segment page. The Segment page consolidates data at the segment level. The upper part of the page includes a trend graph that specifies changes in this segment's member count. Also, if you hover over each data point, it shows the member count for that point. Above the graph, you can find the current member count and last week's growth.
The lower part of the page includes a table with all your segment's members.
The specific fields that appear in this table are based on the attributes of your segment's tables. This table shows only a preview of your records. It presents the first 100 records of your segment so that you can quickly evaluate your segment and go back to the segment editor page to change its definitions.
Recommended segments
Besides the segment builder, another way that you can create segments is through the recommended segments builder. Recommended segments is a quick create option that lets you build simple single operator segments. Segments can be created from:
Profiles - Build a segment based on the unified customer table.
Measures - Build a segment around each of the customer attribute measure types that were previously created.
The New quick segment dialog box guides you through the process to create your segment. The dialog box contains three options:
Field - Defines the attribute that you want to base your segment from.
Operator - Defines the operator that you want to use.
Value - The baseline value that your operator compares against.
Depending on the type of attribute that you select initially, the system provides some other insights that help you create better segments of your customers. For example, if you select a categorical field, the application shows the 10 top customer counts. You're not provided with an operator; you only need to select a value. For a numerical attribute, the system shows what attribute value falls under each customer's percentile. In this case, you would need to choose both an Operator and a Value.
After you save your segment, the system processes the members. After the system finishes processing, you can view your segment like any other segment that you created.