Account-level Audiences

Account-level audiences are audiences for businesses that sell to other businesses. They return the set of accounts which match a combination of account-level traits, user-level traits, and user events. You can sync these accounts and associated users with downstream destinations.

You can use account-level audiences to accomplish the following use cases:

• Identify a set of at-risk accounts based on associated users’ log in patterns, and flag them in your customer service application
• Identify a set of trial accounts which would benefit from a paid plan based on their use, and flag those in your CRM application
• Create an email list of all users in accounts showing high adoption of a specific feature to aid in recruiting for user research

Enable account-level audiences

1. Contact friends@segment.com and provide your workspace ID to have account-level audiences enabled for your workspace. Navigate to Settings > Workspace Settings > General Settings to view your workspace ID.
2. Ensure that group_id is configured as an identifier in Engage Identity Resolution settings. For more information, see Identity Resolution Settings.
3. Instrument group calls to send account information to Segment.

Account-level audience conditions

A single account-level audience can incorporate any combination of the following condition types:

• User-level events (sent through track, page, and screen calls)
• User-level computed and SQL traits
• User-level audience membership
• User-level custom traits (set through an identify call)
• Account-level computed traits and SQL traits
• Account-level audience membership
• Account-level custom traits (set through a group call)

To access account-level audience conditions:

1. Navigate to Engage > Audiences, and click Create.
2. Select Accounts from the Select Type screen.
3. From the Configure screen, select Accounts in the dropdown.

The three types of user-level conditions are:

• Any User (default): Returns all accounts where at least one user associated with the account satisfies the specified condition
• All users: Returns all accounts where all users associated with the account satisfy the specified condition
• None of the users: Returns all accounts where no users associated with the account satisfy the specified condition

Account-level computed and SQL traits

Workspaces with access to account-level audiences can create account-level computed and SQL traits. All user-level computed trait types are supported (see here for a full list). Account-level computed traits operate on the set of events triggered by all users associated with a given account.

Use-cases for account-level computed traits include:

• Calculate the number of times users associated with an account logged in during the past month
• Calculate the average NPS survey score across all NPS surveys submitted by users associated with an account
• Identify the first marketing landing page viewed by any user associated with an account

Use account-level SQL traits to associate users to an account

To associate users to an account with SQL traits, you must return both the group_id and user_id in the account level SQL trait. This fires a Group call which Segment uses to add users to groups in destinations.

When a group (account) contains more than one user, the query returns duplicate group_ids (mapped to unique user_ids). However, Segment doesn’t return duplicate group_ids in the account-level SQL trait. As a result, you can’t map users to accounts in a many-to-many situation.

Use account-level computed and SQL traits as account-level audience conditions

Once created, you can connect account-level computed and SQL traits to downstream destinations. You can also use them as conditions in account-level audiences, enabling you to build audiences based on the set of events triggered by all users associated with a given account.

For example, you can create an audience which selects all accounts where users associated with the account have collectively logged in fewer than 10 times in the past 30 days. To accomplish this, you would:

1. Create an account-level event counter computed trait, selecting the “Logged In” event and specifying a time window of the past 30 days.

   This computed trait will count all “Logged In” events collectively triggered by users associated with each account over the past 30 days, and append the resulting counts to each Account profile.

2. Create an account-level audience, containing a single condition using the new computed trait (for example, logins_past_30_days<10).

Connect account-level audiences, computed traits, and SQL traits to destinations

When you connect an account-level audience or trait to a destination, you select which events to send to the destination when the audience or trait is calculated. Each destination supports a subset of the following events:

• Group: Segment emits one group call per account that matches the conditions. Audience names are included as a boolean trait (for example audience_name:true), while computed and SQL trait values are included directly (for example trait_name:trait_value). An additional group call is emitted when an account no longer matches audience conditions (audience_name:false).
• Identify: Segment emits Identify calls for each user who is associated with any accounts matching the conditions. Audience name is included as a boolean trait (for example audience_name:true), while computed and SQL trait values are included directly (for example trait_name:trait_value). An additional identify call is emitted for each user who is associated with an account which no longer matches audiences conditions (audience_name:false).
• Track: Track calls are an alternative to Identify calls to send events for each user who is associated with any accounts that match the conditions. While identify typically updates a profile, a track event will record an event. The audience name is included as an event property (for example audience_name:true), and the default track name is Audience Entered or Trait Computed for audiences and computed traits, respectively. The event names can be customized. When a user associated with an account which no longer matches audiences conditions, an Audience Exited event is sent with an event property where audience_name:false.

Use account-level traits in user-level audiences

Account-level audiences let you target all users associated with accounts that match your audience criteria. However, you may need to target a subset of users based on the traits of their associated accounts.

Account-level traits are not available in the user-level audience builder. However, account-level audience membership is available in user-level audiences through the “Part of an audience” condition. This enables you to use account-level audiences as a “passthrough” for account-level traits.

For example, you may wish to create an audience which selects all admin-level users associated with accounts in the software industry (for example, user.role=Admin AND account.industry=Software). To accomplish this, you would:

1. Create an account-level audience containing a single account-level custom trait condition that references the account’s industry(industry=Software).
2. Create a user-level audience with two conditions:
   1. A custom trait condition referencing the user’s role (role=Admin)
   2. An audience membership condition referencing the account-level audience you just created (software_industry_audience=True)

Known limitations of account-level audiences

• Unlike user-level audiences, which are computed in real time, account-level audiences are computed on a batched basis. Segment computes account-level audiences roughly every eight hours, but compute times can fluctuate based on system load.
• Account-level audiences don’t respect the context.groupId property on Track calls. If users are associated with multiple accounts (through multiple group calls), the entire collection of a user’s events is considered when evaluating user-level event conditions (not just those events which are tagged with a matching groupId). This can lead to unexpected results where a user’s events triggered in the context of one account lead to another account incorrectly matching an account-level audience.
• The identity breakdown report (displayed in the audience builder for user-level audiences) is not available for account-level audiences.

If you find that these limitations impede your ability to use account-level audiences, contact friends@segment.com with details about your use case.