Segment membership evaluation

Segments are dynamic entities that continuously evolve based on the data sent to the CXP and the filters applied. When new data arrives that matches the segment criteria, the segment is updated. Similarly, objects can leave or join a segment when conditions like rolling date filters are applied, ensuring the segment always reflects the latest data.

This updating process operates as an analytical background task and is independent of the profiling process. As a result, the profiling process remains fast, clear, and free of dependencies on segment updates, ensuring optimal performance and a seamless user experience.

Since segment membership updates are handled as an analytical background process, separate from the profiling process, changes to segment membership are not immediately updated after an event is processed or when relative date filters are applied. This design ensures scalability and efficiency.

#	Trigger Type	What Starts It	What the Segment Evaluator Does	Resulting Activities
1	Event Processing	A new event is ingested (e.g., a purchase).	Evaluates whether the profile now matches or no longer matches existing segment rules.	Creates joined and/or left activities as needed.
2	New Segment Created	User defines a new segment from scratch.	Calculates initial membership "as of now" by evaluating all relevant profiles.	Creates joined activities for all initial members.
3	Segment Updated	Segment rules/filters are changed.	Re-evaluates affected profiles according to the new logic.	Creates joined for newly matching profiles, left for profiles that no longer match.
4	Segment Deleted	A segment is removed.	Treats all current members as leaving the segment.	Creates left activity for all current members.
5	Relative Date Filters	Rolling filters (e.g. "purchased last 30 days").	Precomputes when each profile should enter and leave the segment based on time-based rules (no periodic bulk re-evals).	Schedules or triggers joined and left activities at precomputed times.