The event configuration allows you to define the information Journey Optimizer will receive as events. You can use multiple events (in different steps of a journey) and several journeys can use the same event.
Event configuration is mandatory and must be performed by a data engineer.
You can configure two types of events:
Unitary events: these event are linked to a person. They relate to the behavior of a person (for example, a person bought a product, visited a shop, exited a website, etc.) or something happening linked to a person (for example, a person reached 10 000 loyalty points). This is what Journey Optimizer will listen to in journeys to orchestrate the best next actions. Unitary events can be rule-based or system generated. To learn how to create a unitary event, refer to this page.
Business events: a business event is an event that, in contrast to a unitary event, is not linked to a specific profile. For example, it can be a news alert, a sports update, a flight change or cancellation, an inventory update, weather events, etc. While these events are not specific to a profile, they may be of interest to any number of profiles: individuals subscribed to particular news topics, passengers on a flight, shoppers interested in an out-of-stock product, etc. Business events are always rule-based. When you drop a business event in a journey, it automatically adds a Read segment activity right after. To learn how to create a business event, refer to this page.
If you edit an event used in a draft or live journey, you can only change the name, the description or add payload fields. We strictly limit the edition of draft or live journeys to avoid breaking journeys.
Unitary journeys (starting with an event or a segment qualification) include a guardrail that prevents journeys from being erroneously triggered multiple times for the same event. Profile re-entrance is temporally blocked by default for 5 minutes. For instance, if an event triggers a journey at 12:01 for a specific profile and another one arrives at 12:03 (whether it is the same event or a different one triggering the same journey) that journey will not start again for this profile.
➡️ Discover this feature in video
For business events, the event ID type is always rule-based.
For unitary events, there are two types of event ID:
Rule-based events: this type of event does not generate an eventID. Using the simple expression editor, you simply define a rule which will be used by the system to identify the relevant events that will trigger your journeys. This rule can be based on any field available in the event payload, for example the profile’s location or the number of items added to the profile’s cart.
A capping rule is defined for rule-based events. It limits the number of qualified events that a journey can process to 5000 per seconds for a given Organization. It corresponds to Journey Optimizer SLAs. Refer to your Journey Optimizer licencing and Journey Optimizer Product Description.
System-generated events: these events require an eventID. This eventID field is automatically generated when creating the event. The system pushing the event should not generate an ID, it should pass the one available in the payload preview.
Journey Optimizer requires events to be streamed to Data Collection Core Service (DCCS) to be able to trigger a journey. Events ingested in batch or events from internal Journey Optimizer datasets (Message Feedback, Email Tracking, etc.) cannot be used to trigger a journey. For use cases where you cannot get streamed events, please build a segment based on those events and use the Read Segment activity instead. Segment qualification can technically be used, but can cause downstream challenges based on the actions used. This data does not necessarily need to go to the Real-Time Profile. If you would like to use the events for segmentation or lookup in a separate journey, we recommend you enable the dataset for profile.
Events are POST API calls. Events are sent to Adobe Experience Platform through Streaming Ingestion APIs. The URL destination of events sent through transactional messaging APIs is called an “inlet”. The payload of events follows XDM formatting.
The payload contains information required by Streaming Ingestion APIs to work (in the header) and the information required by Journey Optimizer to work and information to be used in journeys (in the body, for example, the amount of an abandoned cart). There are two modes for the streaming ingestion, authenticated and unauthenticated. For details on Streaming Ingestion APIs, refer to this link.
After arriving through Streaming Ingestion APIs, events flow into an internal service called Pipeline and then in Adobe Experience Platform. If the event schema has the Real-time Customer Profile Service flag enabled and a dataset ID that also has the Real-time Customer Profile flag, it flows into the Real-time Customer Profile Service.
For system-generated events, the Pipeline filters events which have a payload containing Journey Optimizer eventIDs (see the event creation process below) provided by Journey Optimizer and contained in event payload. For rule-based events, the system identifies the event using the eventID condition. These events are listened by Journey Optimizer and the corresponding journey is triggered.
Learn how to configure an event, specify the streaming endpoint and the payload for an event.
Understand the applicable use cases for business events. Learn how to build a journey using a business event and which best practices to apply.