Utilizing the Audiences feature involves several steps to create and send audiences to downstream systems such as Facebook, Yahoo!, Google Analytics, and Nielsen. The high-level steps for defining the audience have been covered in previous articles.
This article focuses on the details of the results produced after the successful execution of audiences. This article covers:
Audience Results Table
The name of the audience table(s) is defined within a data layer. This value is unchangeable after the data layer is created.
While only one value is given, three table names are created based on the value. By default, the syn_audience value is populated while creating the data layer. Following are the names of the tables created based on the syn_audience value given for the field 'Audience Table Name'.
- syn_audience
- syn_audience_metadata
- syn_audience_metrics
In this article, we have used ecom_audience as the value of the 'Audience Table Name'. The following tables will be generated based on that:
- ecom_audience
- ecom_audience_metadata
- ecom_audience_metrics
Previews of these tables can be accessed from:
-
Dataset Screen: Found under the Audience module.
-
Output Dataset Screen: Accessible from any audience group.
- Data Layer >>Audience Screen: Provide additional details such as schema information for audience result tables.
-
Dataset Screen in the Event Store: Provides additional details such as schema information, usage, and more.
Details of these tables are provided below.
-
Results table (syn_audience)
- The results table contains unique identifier values for each audience that match a condition, partitioned by the date the condition was met.
- The example screenshot below shows a preview of data from the Output Datasets screen within the audience group. In the first row, a Visitor with ID 50 met the condition for the attribute "Credit Card-Laptop Purchases" under the attribute group "Payment Methods-Laptop Purchases" on December 4, 2024.
- The table name, which defaults to syn_audience, can only be changed at the time of data layer creation. In the screenshot example below, the table name appears as ecom_audience, reflecting the value provided during the creation of the data layer. This will be the exact name of the table(s) that store the results of all audience builder processes created within that data layer.
- The results table will be created and stored in the specified paths and databases according to the event store selected for the data layer.
Note: The table preview displays results for all audience groups, even when accessing the Output Datasets screen for a specific audience group. This is because all audience groups within a data layer share the same output table. Additionally, the preview provides a limited number of records for reference purposes only. Please refer to the database to view all records in the table. This note applies to all 3 results tables.
-
Metadata (syn_audience_metadata)
- The metadata table contains the unique audience values for all audience builder apps within the data layer. For example, the audience group 'Brand' may have an audience value 'X'.
-
In the example screenshot below, the audience group "Payment Methods-Laptop Purchases" includes an audience named "Credit card-Laptop Purchases. In the second row, this same audience group includes another audience, "Debit Card-Laptop Purchases"
- The tables created for storing metadata are named using the value found in the field 'Audience Table Name' on its Data Layer screen, with "_metadata" added as a suffix. Similar to the results tables, the metadata table is stored in the respective paths and databases defined by the event store for the data layer.
-
Metrics (syn_audience_metrics)
- The metrics table contains statistics on the number of unique identifier values that match each audience condition per partition, typically daily.
- In the example screenshot below, 4 visitors met the condition for the Payment Methods-TV Purchases audience group having the Credit Card - TV Purchases audience on 4th December 2024. In the second row, this same audience group includes another audience, Debit Card-TV Purchases having 3 visitors.
- The tables created for storing metrics are named using the value found in the field 'Audience Table Name' on its Data Layer screen, with "_metrics" added as a suffix. As with the results and metadata tables, metrics tables are stored in the paths and databases specified in the event store for the data layer.
Reviewing Audience Results Under the Data Layer
In the above example, we learned how to preview data under the Output Datasets screen, which is available under any Audience group builder. The same preview section can be accessed on the Datasets screen of the Audience module.
We can access more details about the output datasets like Details, Schema, Preview, and Stats under the Data Layer screen. Follow these steps to navigate to the Audience Dataset screen under Data Layer:
- Open the hamburger menu and click on the
link. This will take you to the Audience module.
- Navigate to the Data Layer screen by clicking 'Data Layer' shown on the left side menu.
- Choose the Data Layer.
- From the left side menu, click on Audiences.
Below is a list of the different properties displayed for each dataset under the data layer:
-
Details & Schema
When viewing from the Data Layer screen, the Details and Schema properties are combined into a single view. The Details section, located at the top of the screen, provides information on several technical aspects of the table. The Schema section, located at the bottom, lists all the columns that exist within the dataset.
-
Preview
As the name suggests, this section provides a preview—displaying a limited number of results from the selected dataset. The data can be sorted and filtered within the application or downloaded as an XLSX file for external review. However, the download is limited to the data shown in the preview screen. The Preview section is also available on every Audience Builder page and Datasets screen under the Audience module.
-
State
This section displays statistical information about the dataset, such as the number of rows and the size of each partition. Partitions are typically organized by date but can also be based on hourly intervals.