Create a User Data Model

User Data Models define the core user dataset that Composable Audiences uses for segmentation. You must create at least one user data model in order to create a composable audience. Like all data models, they are built from a SQL query that selects user records and attributes from your cloud data warehouse, with a required primary ID column to uniquely identify each user. This model forms the foundation for all audiences and can be linked to other models, such as events or subscriptions, for richer targeting.

Example

Here is a simple SQL query you might use to define a user data model for your customers:

SELECT
  user_id,
  email,
  first_name,
  last_name,
  signup_date,
  country,
  marketing_opt_in,
  loyalty_tier
FROM ecommerce.users

• user_id is the primary ID for each record (each user must be unique).
• signup_date helps you identify new versus long-standing users.
• country lets you segment users by geography.
• marketing_opt_in is useful for filtering users who have consented to marketing communications.
• loyalty_tier can be used to group customers into Bronze, Silver, or Gold segments.

With this model in place, you could build an audience such as “new users who signed up in the past 30 days and have opted into marketing emails.”

This example focuses on who the user is (attributes and profile information). When combined with an Event Data Model (which captures what the user does) and a Generic Data Model (which adds extra context), you can build more powerful audiences.

To create a user data model

1. Enter your query

1. Navigate to Data Warehouse > Data Models using the left-hand navigation.
2. Click + Add Data Model in the upper right corner.
3. Select User Data Model.
4. Select a warehouse connection using the Data Connection dropdown menu.
5. Under SQL Query, enter the SQL query you want to execute against your warehouse. You can use joins, filters, grouping, and other SQL capabilities like CASE statements to target the exact data you want to build your audiences with.
6. Click Run Query to ensure that the query runs successfully.
7. Click Next Step to continue.

2. Enter your configuration details

1. Enter a descriptive name for your data model under Model Name.
2. Select a primary ID column (such as user_id or customer_id) retrieved from your SQL query to serve as the unique identifier for each user.

3. Click Save Data Model to continue.

3. Enter your column settings (optional)

1. Use the Enabled toggle to disable or re-enable a column.
2. Review the data type mParticle detects for each column. Update if needed.
3. Enable Autocomplete if you want values from this column to appear as suggestions when entering criteria in the Audience Builder. Values are cached for up to 7 days and then purged.
4. Use the Is Case Sensitive (Yes/No) setting to control casing for autocomplete matches.

Click Save Data Model to continue.

4. Create column mappings

Mappings control how your warehouse data is interpreted in mParticle and by downstream partners.

User identifiers should be mapped to the correct user identifier type (like Email Address or Customer ID) so audiences can be matched and activated within downstream destinations. Descriptive fields, like a loyalty status or signup date, should be mapped to “User Attributes” to make them easier to find and use in the Audience Builder.

Although mappings are optional, you will likely need them to make sure your downstream integrations can effectively use your audience data. Downstream partners usually require identifiers for accurate matching, and mapped attributes make important fields easier for marketers to discover when building marketing campaigns.

• Choose the corresponding field name (user attribute or user identifier).
• Assign a display name for a more marketer-friendly label in the Audience Builder.

5. Confirm your identity settings

After completing your column mapping settings, confirm the identity settings for your user data model.

First, if you want to use this data model for a hybrid audience, you must turn on Enable this Data Model for Hybrid Audiences.

Under Choose identities to match on, review and enable the user identifiers that mParticle should use to match user records from your warehouse with user profiles in mParticle.

The table lists each identifier retrieved from your warehouse and shows:

• The identifier name in your warehouse.
• The corresponding identifier it is mapped to in mParticle.
• The mParticle identifier’s Identity Control configuration.
• The mapping cardinality between your warehouse identifier and your mParticle identifier

In order for the identifiers in your warehouse to appear in this list, they must be correctly mapped to their corresponding mParticle identifiers in the mapping settings.

Identity Control

The Identity Control column shows how the mParticle identifiers that your warehouse identifiers are mapped to are configured for identity resolution.

• Login identity: A login identity can be used to identify one, and only one, MPID for a known user.
• Unique identity: A unique identity value cannot be shared between profiles. For example, if customer_id is a unique identity, two profiles cannot share the same customer ID.
• Immutable identity: An immutable identity cannot be changed once it is set. Immutable identifiers must also be configured as both unique and login identities.

While these are referred to as “Identity Controls,” they are not modifiable here. These are read-only values that reflect how these identifiers are configured in mParticle. You should use this column to help verify that you are mapping your warehouse identifiers to the correct mParticle identifiers

Mapping Cardinality

The Mapping Cardinality column describes how identifiers in your warehouse relate to their corresponding identifiers in mParticle. You may see the following patterns:

• One-to-One: Each warehouse identifier value maps to exactly one mParticle identifier value, and each mParticle identifier value maps to exactly one record in your warehouse. This occurs when:

  • The warehouse identifier is the primary key.
  • The mParticle identifier is configured as unique.

• One-to-Many: Each warehouse identifier value maps to one mParticle identifier value, but the same mParticle identifier value may map to multiple records in your warehouse. This occurs when:

  • The warehouse identifier is the primary key.
  • The mParticle identifier is not unique.

• Many-to-One: Multiple warehouse identifier values can map to the same mParticle identifier value. Each warehouse identifier value still maps to a single mParticle identifier value. This occurs when:

  • The warehouse identifier is not the primary key.
  • The mParticle identifier is unique.

• Many-to-Many: Multiple warehouse identifier values can be mapped between multiple mParticle identifier values. This occurs when:

  • The warehouse identifier is not the primary key.
  • The mParticle identifier is not unique.