Data Ingestion and Connection Configuration
What you must understand
Customer Journey Analytics connections sit on top of Experience Platform datasets, so ingestion questions really test how you map datasets into a connection. Know the difference between a connection (dataset-to-schema mapping, plus the person ID and dataset-level settings) and a data view (the reporting layer). Be able to explain how event datasets differ from profile/lookup datasets, and how lookup datasets are joined to events using a matching key and a defined "keys per match."
Understand the ingestion paths: streaming (Edge Network/Web SDK), batch file uploads, and source connectors. Know that data must first land in Platform with a schema and identity, and that timestamps are critical-CJA uses the event timestamp, and out-of-order or malformed timestamps cause dropped or misattributed rows. Also know the role of backfill/lookback windows and how late-arriving data is handled.
Traps to watch
- Confusing a connection setting with a data view setting. Person ID and dataset inclusion belong to the connection; attribution, session timeout, and components belong to the data view.
- Assuming lookup datasets filter events. They enrich; they do not remove rows unless the join fails and you configure behavior accordingly.
- Forgetting that a dataset can only be used once per connection and that schema changes require care.
- Mixing up the "keys per match" semantics for one-to-many joins.
Rule of thumb
If a question asks where the identity or dataset scope is defined, answer connection. If it asks how a metric is calculated or how sessions break, answer data view.