Design and Publish a Multi-Step Customer Journey

A multi-step customer journey is a stateful, event-driven flow that moves an individual customer through a defined sequence of personalized interactions — receiving a welcome email immediately after account creation, waiting a defined interval before a follow-up, branching on a condition — in response to real-world behavior signals. The core output is a live, published journey that processes incoming events in milliseconds and delivers contextually appropriate actions without manual campaign execution. Practitioners must design the entry condition, the sequence logic, the message content with dynamic personalization, and the journey's timing and capping rules.

Key design decisions include: selecting the right entry event schema (the XDM schema must include the Orchestration eventID field group so the journey engine can match events), deciding between wait-duration and wait-until-event patterns for step sequencing, applying frequency-capping and re-entry rules to prevent message fatigue, and structuring personalization tokens to pull live profile attributes rather than event payload values alone. A well-designed journey also defines an explicit end state and handles journey timeout gracefully.

Parallel viability note (AEP-locked module): Adobe Journey Optimizer's stateful, event-driven orchestration has no clean parallel in composable stacks at equivalent latency. The nearest pattern is an external orchestration engine — Temporal, Apache Airflow, or a custom Lambda/Step Functions workflow — that consumes profile events from a streaming topic, executes step logic, and calls external message APIs. However, these approaches lack the built-in integration with the Real-time Customer Profile for live personalization token resolution, require custom state management for multi-step flows, and cannot match the millisecond event-processing latency of AJO. Teams choosing a composable architecture should plan for significantly higher engineering investment to replicate journey orchestration behavior and should consider AJO as the benchmark for evaluation.