Audiences, Segments, and Profiles
AJO relies on the Real-Time Customer Profile, which unifies identities into a single profile using identity namespaces and merge policies. Audiences (segments) are defined in AEP using the segment builder and can be batch (evaluated on a schedule) or streaming (evaluated as data arrives). AJO journeys can use audience reads as an entry source, and campaigns target audiences directly.
Key concepts: schemas and datasets define the data model; the union schema powers the profile; merge policies determine how conflicting attribute values are resolved across datasets. Segment evaluation methods matter: streaming segments update quickly but have eligibility windows; batch segments update on a schedule. Audience qualification events can trigger journeys when a profile enters or exits a segment. Profile attributes and computed attributes are used for personalization.
The traps: (1) confusing segment (a defined group) with audience in AJO (the terms are often used interchangeably, but in AEP the object is a segment/audience); (2) assuming a streaming segment is instantaneous - there is an evaluation delay; (3) forgetting that merge policy affects which attribute value personalization uses; (4) confusing audience read entry (batch pull) with audience qualification event (real-time trigger); (5) assuming profiles are updated by test mode - they are not.
Worked example
A "high-value customers" segment uses a batch evaluation on order history. A "cart abandoners" segment uses streaming evaluation on a web event. A journey triggered by audience qualification on the cart-abandoner segment fires in near real time; a journey using an audience read on the high-value segment pulls profiles on a schedule.
Rule of thumb
If the journey must react the moment a profile enters a segment, use audience qualification. If it just needs the current membership list on a schedule, use an audience read.