Post Lab Resources

Historical data: When stitching is first enabled, CJA performs an initial backfill for the selected historical period, subject to your package entitlement and the availability of the required persistent/transient ID in the source data. Sources: Stitching Overview, Customer Journey Analytics FAQ

New data: Events are live-stitched going forward as they arrive.

Recently ingested anonymous data: CJA also performs periodic replay stitching, which can reprocess prior events when a user is subsequently identified. Replay only covers the configured lookback window—for example, one day for daily replay or seven days for weekly replay. Data outside that window is not automatically replayed. Sources: Field-based Stitching, Stitching FAQ

Q: How many days does replay stitching look back?

A: Replays can be configured to Daily or Weekly, with options for 14 days or up to 30 days in some CJA product packages.

Sources: Stitching Overview, Field-based Stitching, Stitching FAQ

Data Management and Governance

Q: Will additional derived fields impact processing performance?

A: The addition of derived fields will not impact report processing performance, but highly complex logic in derived fields can impact report response time.

The main performance risks are:

High-cardinality inputs such as full URLs, campaign IDs, or raw query strings.

Large or deeply nested Case When chains.

Heavy or repeated regular expressions.

Multiple Next/Previous or sequential functions.

Lookups using non-normalized keys.

Adobe recommends pushing heavy transformations upstream into Data Prep or lookup datasets, using lower-cardinality normalized fields, and preferring simpler functions such as URL Parse over broad regex logic.

Sources: Derived fields, Derived Fields Guidelines

Q: Are data view changes/admin functions restricted to admins?

A: Data-view visibility is delegated through product profiles, while data-view creation, deletion, and full administrative management are more tightly restricted. Merely being able to use a data view does not grant edit or administrative rights.

Sources: Customer Journey Analytics Access Control, Data Views Overview, Manage Data Views

Comments, Alerts and Annotations

Q: Are Comments available in Adobe Analytics?

A: Annotations and alerts are available in both CJA and AA. Native threaded Workspace comments are currently a CJA-specific capability, not an Adobe Analytics feature.

Sources: Customer Journey Analytics Feature Support, Add And Manage Comments In Projects, Create an annotation, Alerts overview

Q: Where do alerts/notifications appear?

A: Configured Customer Journey Analytics alerts appear in: CJA → Components → Alerts

This opens the Alerts manager, where you can view and manage alerts—edit, enable/disable, copy, tag, renew, or delete them. Alerts created from a Workspace table also appear there. Manage Alerts · Create Alerts

When an alert triggers, its notification is sent to the recipients configured in the alert, such as Adobe users, groups, email addresses, or phone numbers. Notifications can include a link to an automatically generated Analysis Workspace project. Alerts Overview

AI / Coworker

Q: How are AI skills built and maintained?

A: How Skills Are Built

Skills are authored as structured Markdown files (typically SKILL.md) stored under a skills directory. Each skill contains:

Instructions — step-by-step workflows, decision logic, and guardrails the AI follows

API specs — endpoint definitions, parameters, and example payloads

Tool bindings — which MCP tools or API services the skill can call

Dependencies — other skills that must be loaded first

Domain knowledge — product-specific context, rules, and examples

Skills can be created by:

Adobe (platform-provided) — pre-built skills for standard AEP/AJO/CJA workflows (audience creation, journey building, schema exploration, etc.)

Users — custom skills authored conversationally ("save that as a skill", "turn this into a skill") or written directly to the skills directory. These persist across sessions.

How Skills Are Maintained

Versioned on disk — skills live as files that can be edited, replaced, or removed

Loaded on demand — only relevant skills are loaded into context when needed, keeping the system efficient

Dependency-aware — skills declare required and optional dependencies, loaded in order

User-editable — custom skills can be modified or deleted through conversation ("change my X skill", "remove my X skill")

Q: Without a custom skill, would Coworker analyze other customers' data?

A: No. Coworker never accesses another customer's data. Every session is scoped to the authenticated user's IMS organization and sandbox. The system:

Authenticates per-user — all API calls use the current user's token, which limits access to only the orgs and sandboxes that user is entitled to.

Sandbox-isolated — queries, skills, and tools operate within a single sandbox at a time. There is no cross-org or cross-customer data path.

No shared training on customer data — skills are instructions and workflows, not models trained on other customers' content.

Custom skills don't change this boundary — they only add reusable workflows or domain knowledge to your sessions. The underlying data access controls remain the same regardless of whether a skill is custom or platform-provided.

Licensing

Q: How does CJA licensing work relative to data sources?

A: Customer Journey Analytics (CJA) licensing is based on rows of data available for analysis (reportable rows), not on the number or type of data sources you connect. This means you can bring in data from virtually any source — web analytics, mobile, CRM, call center, POS, IoT, offline transactions — without per-source licensing restrictions, as long as your total reportable row volume stays within your contracted entitlement.

"Rows of data" refers to the daily average rows available for analysis across all your connections. Your contract specifies a row entitlement (e.g., 1 million rows). If you exceed it on any given day, prorated overage fees may apply.