Key takeaways

In this module, you learned:

• CJA shifts processing to report time rather than collection time. Traditional Adobe Analytics applies most processing rules (like VISTA rules and processing rules) as data arrives — meaning decisions about what to capture and how to categorize it must be made upfront. CJA is designed for minimal collection-time processing and instead applies powerful transformations at query time through data views and Analysis Workspace, giving analysts far more flexibility after the fact.
• The "person ID" is what allows CJA to unify data across sources into a single journey. When building a connection, you select a person ID field for each dataset. This common identifier is what ties events together — enabling CJA to stitch together touchpoints from different datasets (web, call center, mobile app, in-store) under a single person, which is the foundation for true cross-channel journey analysis.
• A data view is a layer of interpretation on top of a connection. Rather than storing a separate copy of data, a data view is a configuration container that defines how the data from a connection is surfaced in Analysis Workspace — which dimensions and metrics are available, what those components are called, which schema columns they pull from, and how they behave. A single connection can have multiple data views with different configurations.
• CJA stores data in a columnar format for faster analytical queries. Instead of organizing data row by row (as a traditional database might), columnar storage groups all values of each field together. This makes it significantly faster to filter and aggregate large datasets — particularly for the kinds of queries analysts run when slicing data by different dimensions and metrics.
• Custom component IDs keep metrics compatible across different data views. When you assign a custom ID to a metric in a data view, the CJA reporting API uses that ID to identify the metric regardless of which schema field it pulls from. This means a workspace project built with that metric can remain functional even if it is opened against a different data view — an important consideration when building reports intended to be reused across multiple connections or datasets.