Frequently asked questions
蜜豆视频 Experience Platform Segmentation Service provides a user interface and RESTful API that allows you to create audiences through segment definitions or other sources from your Real-Time Customer Profile data. These audiences are centrally configured and maintained on Platform, and are readily accessible by any 蜜豆视频 solution. The following is a list of frequently asked questions regarding audiences and segmentation.
Audience Portal
The following section lists questions related to Audience Portal.
Do I have access to Audience Portal and Audience Composition?
Audience Portal and Audience Composition are available to all Real-Time CDP Prime and Ultimate customers (B2C, B2B, and B2P Editions) and Journey Optimizer Select, Prime, Ultimate Starter, and Ultimate customers.
At this point in time, only profile-based audiences are supported. Support for account-based audiences will be added in a later release.
Are externally generated pre-built audiences supported with Audience Portal?
Yes, externally generated pre-built audiences are supported with Audience Portal. At this point in time, you can import an externally generated audience through a CSV file. In the future, you鈥檒l be able to add audiences through batch or streaming-based source connectors.
What permissions do I need to have in order to upload externally generated audiences?
In order to upload externally generated audiences, you need to have the 鈥淰iew segments鈥, 鈥淢anage segments鈥, and 鈥淚mport audiences鈥 permissions. There are no specific role-based controls required to upload externally generated audiences.
What happens when I upload an externally generated audience?
When you upload an externally generated audience, a dataset will be created and be visible within the dataset inventory. The name of the dataset will be the same as the name of the externally generated audience you uploaded.
What is an externally generated audience comprised of, and what happens to this data when it鈥檚 imported to Platform?
During the import external audience workflow, you must specify which column in the CSV file corresponds with the Primary Identity. An example of a primary identity includes email address, ECID, or an organization-specific custom identity namespace.
The data associated with this primary identity column is the only data that is attached to the profile. If there are no existing profiles that match the data in the primary identity column, a new profile is created. However, this profile is essentially an orphaned profile since no attributes or experience events are associated with this profile.
All the other data within the externally generated audience are considered payload attributes. These attributes can only be used for personalization and enrichment during activation, and are not attached to a profile. These attributes are, however, stored in the data lake.
While the externally generated audience can be referenced when creating audiences using the Segment Builder, individual profile attributes cannot be used.
Can I reconcile externally generated audience data with an existing profile in Platform?
Yes, the externally generated audience will be merged with the existing profile in Platform if the primary identifiers match.This data can take up to 24 hours to be reconciled. If profile data does not already exist, a new profile will be created as the data is ingested.
How are customer consent preferences honored for externally generated audiences that are imported into Audience Portal? consent
As customer data is captured from multiple channels, identity stitching and merge policies allow this data to be consolidated in a single Real-Time Customer Profile. Information on the customers鈥 consent preferences are stored and evaluated at the profile level.
Downstream destinations check each profile for consent information prior to activation. Each profile鈥檚 consent information is compared against consent requirements for a particular destination. If the profile does not satisfy the requirements, that profile is not sent to a destination.
When an external audience is ingested into Audience Portal, they are joined with existing profiles using a primary ID such as email or ECID. As a result, the existing consent policies will remain in force throughout activation.
Please note you should not include consent information with an externally generated audiences, since the payload variables are not stored in the Profile store but in the data lake. Instead, you must use an 蜜豆视频 Experience Platform ingestion channels where profile data is imported.
Can I use an externally generated audience to build other audiences?
Yes, any externally generated audience will appear within the audience inventory and can be used when building audiences within the Segment Builder.
How often are externally generated audiences evaluated?
Externally generated audiences are only evaluated during the time of import. Since the associated attributes to these import audiences are non-durable and are not part of the Profile store, the only time an externally generated audience will be updated is if the existing audience is manually updated.
Can I use externally uploaded attributes as part of segmentation?
No, you cannot. Profile attributes are meant to be long-lasting attributes, while externally generated audience data that is uploaded only contains contextual data that is associated with that externally generated audience.
The externally generated audience鈥檚 contextual data, or enrichment attributes, are not durably long lasting, as their lifecycle is tied to the uploaded audience. As a result, due to its transient nature, these enrichment attributes are not available for use in segmentation.
However when mapping your audiences to batch or file-based destinations, you can use these externally generated enrichment attributes to augment your audiences and further downstream activations.
To learn more about this capability, please read the guide on activating audience data to batch profile export destinations.
Is there a specific merge policy for externally generated audiences?
The organization-specific default merge policy is automatically applied when uploading externally generated audiences. However, you can change the merge policy that is applied to the externally generated audience during the import audience workflow.
Where can I activate externally generated audiences to?
An externally generated audience can be mapped to any destination and can be used in 蜜豆视频 Journey Optimizer campaigns.
Can I delete an externally generated audience?
Yes! Externally generated audiences can be deleted within Audience Portal.
What should I do if I accidentally uploaded an externally generated audience?
If you have accidentally uploaded an externally generated audience and you want to remove the data, you can clear the profiles associated with the audience by uploading a CSV file with one row and no data.
How long do externally generated audiences last for?
The current data expiration for externally generated audiences is 30 days. This data expiration was chosen to reduce the amount of excess data stored within your organization.
After the data expiration period passes, the associated dataset will still be visible within the dataset inventory, but you will not be able to activate the audience and the profile count will show as zero.
Is there a maximum number of externally generated audiences I can import?
There is no limit to the number of externally generated audiences you can import. However, please note that the imported audiences do count against the overall audience limit.
How will Audience Portal and Audience Composition interact with the release of Real-Time CDP Partner Data?
Audience Portal and Audience Composition will interact with Partner Data in two ways:
- If you ingest a partner-provided prospect list using the Prospect Profile class and workflow, the prospects will be kept separately from merge customer profiles in Profile Service. As a result, this means that prospect lists will not appear in either Audience Portal or Audience Composition for use.
- If you are leveraging partner-provided attributes to enrich existing first-party profiles, those partner-data-enriched audiences will appear in both Audience Portal and Audience Composition for use.
How can I use additional attributes with my audiences?
With audiences, there are two different types of additional attributes you can add - payload (contextual) attributes and enrichment attributes.
Payload attributes are attributes that are ingested as part of the CSV upload of an externally generated audience. These attributes are not ingested into the Real-Time Customer Profile, but can be used as a part of a downstream destination.
Enrichment attributes are attributes that come from a dataset and are joined with an audience in Audience Composition. These attributes can currently only be used in 蜜豆视频 Journey Optimizer campaigns. Support for 蜜豆视频 Journey Optimizer journeys is coming soon, with support for downstream destinations pending future release.
Lifecycle states lifecycle-states
The following section lists questions related to lifecycle states and lifecycle state management within the Audience Portal.
What do the different lifecycle states represent?
The following chart explains the different lifecycle statuses, what they represent, where audiences with that status can be used, as well as impact on segmentation guardrails.
In what states can I edit my audiences in?
Audiences can be edited in the following lifecycle states:
- Draft: If an audience is edited in the draft state, it will remain in the draft state unless it is explicitly published.
- Published: If an audience is edited in the published state, it will remain published, and the audience will be automatically updated.
- Inactive: If an audience is edited in the inactive state, it will remain inactive. This means that it will not be evaluated or updated. If you need to update the audience, you鈥檒l need to publish the audience.
Once an audience is deleted, it cannot be edited.
What lifecycle states can I move an audience to?
The possible lifecycle states that an audience can be moved to depends on the current state of the audience.
If your audience is in the draft state, you can either publish or delete it if the audience does not have any dependents.
If your audience is in the published state, you can either deactivate or delete it if the audience does not have any dependents.
If your audience is in the inactive state, you can either re-publish or delete it if the audience does not have any dependents.
Are there any caveats for audiences in certain lifecycle states?
Audiences in the published state can only be moved to another state if the audience does not have any dependents. This means that if your audience is used in a downstream service, it cannot be deactivated or deleted.
If an audience that is evaluated using batch segmentation is re-published, which is when an audience goes from inactive to published, the audience will refresh after the daily batch job. When it is first re-published, the profiles and data will be the same as when the audience was made inactive.
How do I put an audience in the draft state?
The method to put an audience into the draft state depends on the origin of the audience.
For audiences created using Segment Builder, you can set the audience to the draft state by selecting 鈥淪ave as draft鈥 in Segment Builder.
For audiences created in Audience Composition, audiences are automatically saved as a draft until published.
For audiences that are externally created, audiences are automatically published.
Once an audience is in the published state, you cannot change the original audience back into the draft state. However, if you copy the audience, the newly copied audience will be in the draft state.
How do I put an audience in the published state?
For audiences created using Segment Builder or Audience Composition, you can set the audience to the published state by selecting 鈥淧ublish鈥 in their respective UIs.
Audiences that are externally created are automatically set to published.
How do I put an audience in the inactive state?
You can put a published audience into the inactive state by opening the quick actions menu in Audience Portal and selecting 鈥淒eactivate鈥.
How do I re-publish an audience?
You can re-publish an audience by selecting an audience that is in the inactive state, opening the quick actions menu on Audience Portal and selecting Publish.
How do I put an audience in the deleted state?
You can put an audience into the delete state by opening the quick actions menu in Audience Portal and selecting Delete.
Are there any caveats for lifecycle state transitions?
Yes, there are some caveats to be aware of when you are using audiences in downstream services such as 蜜豆视频 Journey Optimizer or non-customer-based audiences such as account-based audiences.
At this time, you must manually check if the audience is used downstream in 蜜豆视频 Journey Optimizer, as this status is currently not automatically checked.
Additionally, you must manually check if the audience is used as a component of an account-based audience, as this status is also not currently automatically checked.
What happens when I copy an audience? copy
When you copy an audience, the new audience will be in the draft state, and retain the same folders, tags, and labels that were applied to the original audience.
Does using an audience as a child audience affect lifecycle state transitions?
Yes, using an audience as a child audience does affect what lifecycle states transitions the child and parent audience can undertake.
In order for a child audience to be moved to the published state, all of its parent audience must be in the published state. Parent audiences can either be published before publishing the child audience or, if the user confirms, can be automatically published when the child audience is published.
In order for the parent audience to be moved to the inactive or deleted state, all of its child audiences must be deactivated or deleted.
Can I refer to an audience that is in a different lifecycle state?
Yes! If your audience is currently in the draft state, you can refer to audiences in either the draft or published state. However, in order to publish this audience, you must publish the other parent audiences.
Audience inventory
The following section lists questions related to audience inventory within the Audience Portal.
Do I need additional permissions to use audience inventory features?
No, you do not. So long as you have edit permissions for audiences, you鈥檒l be able to create, update, and manage your folders and tags within the Audience Portal. For more information about managing permissions, please read the manage permissions guide.
Is there a limit to the number of folders I can create?
No, there is no limit to the number of folders you can create. For more information on folders, please read the audience inventory section of the Segmentation Service UI overview.
Is there a limit to the number of tags that can be added to an audience?
No, there is no limit to the number of tags that can be added to an audience. For more information on tags, please read the audience inventory section of the Segmentation Service UI overview.
Is there a limit to the number of tags I can create?
No, there is no limit to the number of tags that you can create. However, you can create a maximum of 100 categories to apply for the tags. For more information about tag management, please read the Managing Tags guide.
When I search for an audience by name or tag in a parent folder, can I also search through the related child folders?
No, this behavior is not supported. However, you can change the audience inventory view to look at All Audiences, then search across all the folders. For more information on using search in audience inventory, please read the search section of the Segmentation Service UI overview.
Can I automatically assign an audience into a folder at the time of creation?
At this point in time, no. However, this capability may be available in the future.
Can I move multiple audiences into a folder at the same time?
At this point in time, no. However, this capability may be available in the future.
Audience Composition
The following section lists questions related to Audience Composition.
When should I use Audience Composition as opposed to using the Segment Builder?
Both Audience Composition and Segment Builder have important roles in the creation of building audiences in Platform.
The Segment Builder is more suited for audience creation (for building an audience from scratch), while Audience Composition is more suited for audience curation and personalization (for creating new audiences based on an existing audience).
The following table illustrates the difference between the two services:
- Single stage audience generation
- Creates the basic blocks of audiences from profile, time-series, and multi-entity data
- Used to create one audience
- Multi-stage audience generation, using set based operations
- Uses the audiences created by the Segment Builder and applies data enrichment options such as ranking profile attributes and splitting into sub-audiences
- Used to create multiple audiences at once
To learn more about the Segment Builder, please read the Segment Builder guide. To learn more about Audience Composition, please read the Audience Composition guide.
Can I use externally generated audiences in Audience Composition?
At this point in time, no. However, this capability should be available in the near future.
Can I send audiences from Audience Composition to all downstream destinations and channels?
At this point in time, no. Currently, you can use audiences from Audience Composition in 蜜豆视频 Journey Optimizer Campaigns and Real-Time CDP destinations. 蜜豆视频 Journey Optimizer Journeys will be supported in a future release.
Are there any guardrails on the number of compositions?
At this point in time, you can only have 10 published compositions per sandbox. This guardrail is planned to be increased in a future release.
What are the workflow guardrails for Audience Composition?
The composition component placing follows a rigid structure as follows:
- You always start with the Audience block to select your starting activity. You can have a maximum of one Audience block.
- You can optionally add an Exclude block that follows the Audience block.
- You can optionally add an Enrich block that follows the Exclude block. You can only use one Enrich block per composition.
- You can optionally add a Rank or Split block. You can only have one of these blocks per composition.
- You always end with a Save block to save your audience.
Additionally, the following restrictions(?) apply when using these blocks:
-
Split block
- This block only supports String data types. The Split block does not support the date or boolean data type.
- Additionally, this block does not support enrichment attributes.
-
Exclude block
- This block does not support the date or boolean data type.
-
Rank block
- This block does not support enrichment attributes.
For more details about using Audience Composition, please read the Audience Composition UI guide.
When are audiences created using Audience Composition saved and evaluated?
Audiences are automatically saved while creating them in Audience Composition. The audience鈥檚 creation time will be the first time this automatic save occurs.
After the audience composition has been created, it can take up to 48 hours for it to be evaluated and activated for usage in downstream services such as a Real-Time CDP destination or 蜜豆视频 Journey Optimizer channel.
When can I use the audience I created?
The audience created in Audience Composition will immediately show up in Audience Portal. However, in order to use it in 蜜豆视频 Journey Optimizer, you must wait at least 24 hours after evaluation.
Are evaluation jobs visible within the monitoring section?
At this time, evaluation jobs are not displayed within the monitoring UI.
Can I use an Audience Composition in another composition?
No, audiences created using Audience Composition cannot be used as an input in another audience composition.
How does splitting work in Audience Composition?
Audience splitting lets you further subset your audience into smaller groups.
When splitting by attribute, there is mutual exclusivity between the groups. This means that if a record meets the criteria of multiple split paths, it will be assigned the first path from the left and not assigned to any of the other paths.
When splitting by percentage, splits are randomly done. This means that the profiles will be randomly assigned to each path.
For more information on the Split block, please read the Audience Composition UI guide.
Can I use all segmentation types in the Audience Composition workflow?
Yes, all segmentation types (batch segmentation, streaming segmentation, and edge segmentation) are supported in the Audience Composition workflow. However, since compositions are currently only run once per day, even if streaming- or edge-evaluated audiences are included, the result will be based on audience membership at the time the composition was executed.
Audience membership
The following section lists questions related to audience membership.
How can I confirm a profile鈥檚 membership in an audience?
To confirm a profile鈥檚 audience membership, visit the profile details page of the profile you want to confirm. Select Attributes, followed by View JSON, and you can confirm that the segmentMembership object contains the ID of the audience.
How does batch segmentation resolve profile membership?
Audiences evaluated using batch segmentation resolve daily, with the audience membership results being recorded into the profile鈥檚 segmentMembership attribute. Profile lookups generate a fresh version of the profile at the time of the lookup, but it does not refresh the batch segmentation results.
As a result, when changes are made to the profile, such as merging two profiles together, these changes will appear in the profile when looked up, but will not be reflected in the segmentMembership attribute until the segment evaluation job has run again.
For example, let鈥檚 say you鈥檝e created two mutually exclusive audiences: Audience A is for people who live in Washington and Audience B is for people who do not live in Washington. There are two profiles - profile 1 for a person who lives in Washington, and profile 2 for a person who lives in Oregon.
When the batch segmentation evaluation job runs, profile 1 will go to Audience A, while profile 2 will go to Audience B. Later on, but before the next day鈥檚 batch segmentation evaluation job runs, an event that reconciles the two profiles enters Platform. As a result, a single merged profile that contains profiles 1 and 2 is created.
Until the next batch segment evaluation job is run, the new merged profile will have audience membership in both profile 1 and profile 2. As a result, this means it鈥檒l be a member of both Audience A and Audience B, despite the fact that these audiences have contradictory definitions. For the end-user, this is the exact same situation as before the profiles were connected, since there was always just the one person involved, and Platform just did not have enough information to connect the two profiles together.
If you use profile lookup to retrieve the newly created profile and look at its audience membership, it鈥檒l show that it鈥檚 a member of both Audience A and Audience B, despite the fact that both of these audiences have contradictory definitions. Once the daily batch segmentation evaluation job runs, the audience membership will be updated to reflect this updated state of profile data.
If you need more real-time audience resolution, use streaming or edge segmentation.
How long does it take for streaming data to be available in batch segmentation workflows?
It may take up to three hours for streaming data to be available in batch segmentation workflows.
For example, if a batch segmentation job runs at 9PM, it is guaranteed to contain streaming ingested data up until 6PM. Streaming ingested data that was ingested after 6PM but before 9PM may be included.