Audiences and Segmentation
This domain tests how you turn raw events and attributes into audience membership: criteria logic, real-time versus standard evaluation, and how membership leaves mParticle.
Building blocks
Audiences are built from criteria groups joined by AND/OR. A condition can test a user attribute, an event performed or not performed within a time window, or membership in another audience. Membership is stored on the user profile and can be sent to integrations through an audience connection.
Real-time vs standard
- Real-time audiences are evaluated as each event arrives. They suit suppression, triggered messaging, and short windows, but cannot perform long historical lookback.
- Standard (batch) audiences recalculate on a schedule and support historical conditions such as "did not purchase in the last 30 days."
- Nested audiences propagate membership, so a parent audience can lag the child.
Worked example
"Lapsed high-value": lifetime purchase value at least 500 AND no purchase in the last 30 days AND member of VIP. The lookback rules out real-time, so build a standard audience. Add VIP as a nested criterion rather than duplicating its logic. Then connect the audience to the email integration as an audience input.
Traps
- Choosing real-time for a lookback condition it cannot evaluate.
- Assuming removal from an audience reaches every partner instantly.
- Referencing attributes or events that are absent from the data plan.
- Misreading AND/OR precedence across multiple criteria groups, which widens or narrows the segment unexpectedly.