Adobe Experience Platform is an open system that transforms your data into robust customer profiles that update in real time and uses AI-driven insights to help you to deliver the right experiences across every channel. You can ingress data of varying types, volumes, and histories to Experience Platform using sources and then cater that data to use cases ranging from segmentation and personalization to analytics and machine learning.
Platform offers licenses that establish the number of profiles that you can create and the amount of data that you can bring in. Given the capacity to bring in any source, volume, or history of data, it is possible to exceed your licensing entitlements as your data volumes grow.
This document outlines best practices to follow and tools you can use to better manage your license entitlements with Adobe Experience Platform.
Experience Platform is primarily composed of two data repositories: the data lake and the Profile store.
The data lake primarily serves the following purposes:
The Profile store is where customer profiles are created and primarily serves the following purposes:
Your access to the data lake can depend on the product SKU that you purchased. For more information on product SKUs, please speak with your Adobe representative.
When you license Experience Platform, you are provided with license usage entitlements that vary depending on SKU:
Addressable Audience - the total number of customer profiles that are contractually allowed in Experience Platform, including both known and pseudonymous profiles.
Total Data Volume - the total amount of data available for Adobe Experience Platform Profile Service to use in engagement workflows.
The availability of these metrics and the specific definition of each of these metrics varies depending on the licensing that your organization has purchased.
The Adobe Experience Platform UI provides a dashboard through which you can view a snapshot of your organization’s license-related data for Platform. The data in the dashboard is displayed exactly as it appears at the specific point in time when the snapshot was taken. The snapshot is neither an approximation nor a sample of data, and the dashboard is not updating in real-time.
For more information, see the guide on using the license usage dashboard on Platform UI.
The following sections outline best practices to follow to better manage your data.
Not all data is the same in Adobe Experience Platform. Some data may be dense, but low in value, while others may be sparse, but high in value. Some data may lose value as soon as its generated, while others may be valuable for months, if not years.
There are three dimensions to consider in understanding the value of your data:
Dimension | Description | Example |
---|---|---|
Volume | Represents the amount and totality of data ingested. | Web clicks - high in volume and moderate in fidelity. Value may diminish quickly. |
Timespan | Represents the length of time that ingested data continues to stay valuable. | Offline purchases - moderate in volume and fidelity, but may be valuable for long periods of time. |
Fidelity | Represents how rich the data is with information. | Customer accounts - low in volume, but high in fidelity. Can be valuable beyond the lifetime of a customer. |
There are two central scenarios to consider when ensuring that your data usage remains within your license entitlement limits:
Data can be ingested into one or multiple systems in Platform, namely the data lake and/or the Profile store. This means that different data can exist in both systems for a variety of different use cases. For example, you may want to hold historical data in the data lake, but not in the Profile store. You can select which data to send to the Profile store by enabling a dataset for Profile ingestion.
Your access to the data lake can depend on the product SKU that you purchased. For more information on product SKUs, please speak with your Adobe representative.
You can apply both data ingestion filters and expiration rules to remove data that has become obsolete for your use cases. Typically, behavioral data (such as Analytics data) consumes significantly more storage than record data (such as CRM data). For example, many Platform users have upwards of up to 90% of profiles being populated by behavioral data alone, in comparison to that of record data. Therefore, managing your behavioral data is critical in ensuring compliance within your license entitlements.
There are a number of tools that you can leverage to stay within your license usage entitlements:
Identity graphs do not count towards your total addressable audience entitlement because addressable audience refers to your total count of customer profiles.
However, identity graph limits can affect your addressable audience due to splitting identities. For example, if the oldest ECID is removed from the graph, ECID will continue to exist in Real-Time Customer Profile as a pseudonymous profile. You can set Pseudonymous profile data expirations to circumvent this behavior. For more information, read the guardrails for Identity Service data.
Ingestion filters allow you to bring in only the data that is needed for your use cases and filters out all events that are not required.
Ingestion filter | Description |
---|---|
Adobe Audience Manager source filtering | When you create an Adobe Audience Manager source connection, you can pick and choose which segments and traits to bring into the data lake and Real-Time Customer Profile, rather than ingesting the Audience Manager data in its entirety. See the guide on creating an Audience Manager source connection for more information. |
Adobe Analytics Data Prep | You can use Data Prep functionalities when creating an Analytics source connection to filter out data that is not required for your use cases. Through Data Prep, you can define which attributes/columns need to be published to Profile. You can also provide conditional statements to inform Platform whether data is expected to be published to Profile, or just to the data lake. See the guide on creating an Analytics source connection for more information. |
Support for enable/disable datasets for Profile | To ingest data into the Real-Time Customer Profile, you must enable a dataset for use in the Profile store. Doing so, adds to your Addressable Audience and Total Data Volume entitlements. Once a dataset is no longer required for customer profile use cases, you can disable that dataset’s integration to Profile to ensure that your data remains license compliant. See the guide on enabling and disabling datasets for Profile for more information. |
Web SDK and Mobile SDK data exclusion | There are two types of data collected by Web and Mobile SDK: data that is collected automatically and data that is explicitly collected by your developer. To better manage license compliance, you can disable automatic data collection in the configuration of the SDK through the context setting. Custom data can also be removed or not set by your developer. |
Server-side forwarding data exclusion | If you are sending data to Platform using server-side forwarding, you can exclude what data is sent by either removing the mapping in a rule action to exclude it across all events, or by adding conditions to the rule so that data only fires for certain events. See the documentation on events and conditions for more information. |
Filter data at the source level | You can use logical and comparison operators to filter row-level data from your sources before creating a connection and ingesting data to Experience Platform. For more information, read the guide on filtering row-level data for a source using the Flow Service API. |
The Profile store is composed of the following components:
Profile store component | Description |
---|---|
Profile fragments | Each customer profile is composed of multiple profile fragments that have been merged to form a single view of that customer. For example, if a customer interacts with your brand across several channels, your organization will have multiple profile fragments related to that single customer appearing in multiple datasets. When these fragments are ingested into Platform, they are stitched together using the identity graph to create a single profile for that customer. Profile fragments consist of an identity namespace as the identifier, with associated record data and/or time-series data. |
Record data (Attributes) | A profile is a representation of a subject, an organization or an individual, composed of many Attributes (also known as record data). For example, the profile of a product may include a SKU and description, whereas the profile of a person contains information like first name, last name, and email address. Record data is usually low/moderate in volume, but valuable for long periods of time. |
Time-series data (Behavior) | Time-series data provides information about a user behavior. Represented by the standard schema class Experience Data Model (XDM) ExperienceEvent, time-series data can describe events such as items being added to a cart, links being clicked, and videos viewed. The value of behavioral may diminish over time. |
Identity namespace (identities) | As customer data comes together, it is merged into a single profile through the use of identity namespaces, and the ability to stich these identities together as more information becomes known about the user. See the identity namespaces overview for more information. |
There are a number of reports available to help you understand the composition of the Profile store. These reports help you make informed decisions about how and where to set your Experience Event expirations to better optimize your license usage:
This capability allows you to automatically remove stale Pseudonymous Profiles from the Profile store. For more information on this feature, please read the Pseudonymous Profile data expiration overview.
This capability allows you to automatically remove behavioral data from a Profile-enabled dataset that is no longer valuable for your use cases. See the overview on Experience Event expirations for details on how this process works once it is enabled for a dataset.
The following is a list of some recommended best practices that you can follow to ensure better adherence to your license usage entitlement:
The best practices and tools outlined in this document will help you better manage your license entitlement usage within Adobe Experience Platform. This document will be updated as additional features are released to help provide visibility and control to all Experience Platform customers.
The following table outlines the list of currently available features at your disposal, to better manage your license usage entitlement.
Feature | Description |
---|---|
Enable/Disable Datasets for Profile | Enable or disable dataset ingestion into Real-Time Customer Profile. |
Experience Event expirations | Apply an expiration time for all events ingested into a Profile-enabled dataset. Please contact your Adobe account team or Customer Care to enable this feature. |
Adobe Analytics Data Prep filters | Apply Kafka filters to exclude unnecessary data from ingestion |
Adobe Audience Manager source connector filters | Apply Audience Manager source connection filters to exclude unnecessary data from ingestion |
Event forwarding data filters | Apply server-side Kafka filters to exclude unnecessary data from ingestion. See the documentation on tag rules for additional information. |
License Usage Dashboard UI | View a snapshot of your organization’s license-related data for Experience Platform |
Dataset Overlap Report API | Outputs the datasets that contributes the most to your Addressable Audience |
Identity Overlap Report API | Outputs the identity namespaces that contribute the most to your Addressable Audience |