Show Menu

Prepare data for use in Intelligent Services

In order for Intelligent Services to discover insights from your marketing events data, the data must be semantically enriched and maintained in a standard structure. Intelligent Services leverage Experience Data Model (XDM) schemas in order to achieve this. Specifically, all datasets that are used in Intelligent Services must conform to the Consumer ExperienceEvent (CEE) XDM schema.
This document provides general guidance on mapping your marketing events data from multiple channels to this schema, outlining information on important fields within the schema to help you determine how to effectively map your data to its structure.

Workflow summary

The preparation process varies depending on whether your data is stored in Adobe Experience Platform or externally. This section summarizes the necessary steps you need to take, given either scenario.

External data preparation

If your data is stored outside of Experience Platform, follow the steps below:
  1. Contact Adobe Consulting Services to request access credentials for a dedicated Azure Blob Storage container.
  2. Using your access credentials, upload your data to the Blob container.
  3. Work with Adobe Consulting Services get your data mapped to the Consumer ExperienceEvent schema and ingested into Intelligent Services.

Experience Platform data preparation

If your data is already stored in Platform, follow the steps below:
  1. Review the structure of the Consumer ExperienceEvent schema and determine whether your data can be mapped to its fields.
  2. Contact Adobe Consulting Services to help map your data to the schema and ingest it into Intelligent Services, or follow the steps in this guide if you want to map the data yourself.

Understanding the CEE schema

The Consumer ExperienceEvent schema describes the behavior of an individual as it relates to digital marketing events (web or mobile) as well as online or offline commerce activity. The use of this schema is required for Intelligent Services because of its semantically well-defined fields (columns), avoiding any unknown names that would otherwise make the data less clear.
The CEE schema, like all XDM ExperienceEvent schemas, captures the time-series-based state of the system when an event (or set of events) occurred, including the point in time and the identity of the subject involved. Experience Events are fact records of what occurred, and thus they are immutable and represent what happened without aggregation or interpretation.
Intelligent Services utilize several key fields within this schema to generate insights from your marketing events data, all of which can be found at the root level and expanded to show their required subfields.
Like all XDM schemas, the CEE mixin is extensible. In other words, additional fields can be added to the CEE mixin, and different variations can be included in multiple schemas if required.
A complete example of the mixin can be found in the public XDM repository . In addition, you can view and copy the following JSON file for an example of how data can be structured to comply with the CEE schema. Refer to both of these examples as you learn about the key fields outlined in the section below, in order to determine how you can map your own data to the schema.

Key fields

There are several key fields within the CEE mixin which should be utilized in order for Intelligent Services to generate useful insights. This section describes the use case and expected data for these fields, and provides links to reference documentation for further examples.

Mandatory fields

While the use of all key fields is strongly recommended, there are two fields that are required in order for Intelligent Services to work:

Primary identity

One of the fields in your schema must be set as a primary identity field, which allows Intelligent Services to link each instance of time-series data to an individual person.
You must determine the best field to use as a primary identity based on the source and nature of your data. An identity field must include an identity namespace that indicates the type of identity data the field expects as a value. Some valid namespace values include:
  • "email"
  • "phone"
  • "mcid" (for Adobe Audience Manager IDs)
  • "aaid" (for Adobe Analytics IDs)
If you are unsure which field you should use as a primary identity, contact Adobe Consulting Services to determine the best solution.


This field represents the datetime at which the event occurred. This value must be provided as a string, as per the ISO 8601 standard.


This field is only mandatory when using Attribution AI.
This field represents the marketing channel related to the ExperienceEvent. The field includes information about the channel type, media type, and location type.
Example schema
  "@id": "",
  "@type": "",
  "xdm:mediaType": "earned",
  "xdm:mediaAction": "clicks"

For complete information regarding each of the required sub-fields for xdm:channel , please refer to the experience channel schema spec. For some example mappings, see the table below .
Example channel mappings
The following table provides some examples of marketing channels mapped to the xdm:channel schema:
Paid Search
https:/ /
Social - Marketing
https:/ /
https:/ /
https:/ /
Internal Referrer
https:/ /
Display ViewThrough
https:/ /
QR Code Redirect
https:/ /
https:/ /

Mapping and ingesting data

Once you have determined whether your marketing events data can be mapped to the CEE schema, the next step is to determine which data you to bring into Intelligent Services. All historical data used in Intelligent Services must fall within the minimum time window of four months of data, plus the number of days intended as a lookback period.
After deciding the range of data you want to send, contact Adobe Consulting Services to help map your data to the schema and ingest it into the service.
If you have an Adobe Experience Platform subscription and want to map and ingest the data yourself, follow the steps outlined in the section below.

Using Adobe Experience Platform

The steps below require a subscription to Experience Platform. If you do not have access to Platform, skip ahead to the next steps section.
This section outlines the workflow for mapping and ingesting data into Experience Platform for use in Intelligent Services, including links to tutorials for detailed steps.

Create a CEE schema and dataset

When you are ready to start preparing your data for ingestion, the first step is to create a new XDM schema that employs the CEE mixin. The following tutorials walk through the process of creating a new schema in the UI or API:
The tutorials above follow a generic workflow for creating a schema. When choosing a class for the schema, you must use the XDM ExperienceEvent class . Once this class has been chosen, you can then add the CEE mixin to the schema.
After adding the CEE mixin to the schema, you can add other mixins as required for additional fields within your data.
Once you have created and saved the schema, you can create a new dataset based on that schema. The following tutorials walk through the process of creating a new dataset in the UI or API:
After the dataset is created, you can find it in the Platform UI within the Datasets workspace.

Add identity fields to the dataset

Future releases of Intelligent Services will integrate Adobe Experience Platform Identity Service into their customer identification capabilities. As such, the steps outlined below are subject to change.
If you are bringing in data from Adobe Audience Manager, Adobe Analytics, or another external source, then you have the option to set a schema field as an identity field. To set a schema field as an identity field, view the section on setting identity fields within the UI tutorial for creating a schema using the Schema Editor or alternatively the API tutorial .
If you are ingesting data from a local CSV file, you can skip ahead to the next section on mapping and ingesting data .

Map and ingest data

After creating a CEE schema and dataset, you can start mapping your data tables to the schema and ingest that data into Platform. See the tutorial on mapping a CSV file to an XDM schema for steps on how to perform this in the UI. You can use the following sample JSON file to test the ingestion process before using your own data.
Once a dataset has been populated, the same dataset can be used to ingest additional data files.
If your data is stored in a supported third-party application, you can also choose to create a source connector to ingest your marketing events data into Platform in real time.

Next steps

This document provided general guidance on preparing your data for use in Intelligent Services. If you require additional consulting based on your use case, please contact Adobe Consulting Support.
Once you have successfully populated a dataset with your customer experience data, you can use Intelligent Services to generate insights. Refer to the following documents to get started: