CJA connections and data views: the two setup steps people mix up

Why this matters

This distinction matters because mixing up a connection and a data view leads to misaligned analysis. If a team builds an analysis project on a connection instead of a data view, they miss critical configuration like field definitions, persistence, or attribution rules—resulting in inaccurate insights. Without proper data view setup, even if datasets are connected, the data won’t behave as intended across channels or over time.

Also, assuming all data is available without defining a data view means teams can’t isolate or compare specific user behaviors. For example, a marketing team might see product team metrics if they don’t have a separate data view with correct field settings. This confusion leads to flawed decisions, especially when cross-channel analysis is needed. Clear separation ensures each team sees only what they need, with accurate, reliable data.

The key ideas

In Customer Journey Analytics, a connection brings one or more datasets from Experience Platform into CJA. These datasets must all include a Person ID — the field that links each row to a specific person. When datasets from different channels share the same Person ID, they can be combined in one connection, enabling cross-channel analysis. This means you can see how a customer interacts across email, web, and app over time.

A data view sits on top of a connection and defines how the data is structured for analysis. It chooses which fields become dimensions or metrics and sets rules like how long data persists or how attribution is calculated. Multiple data views can use the same connection — for example, one for marketing and another for product teams — allowing different groups to analyze data in tailored ways.

Analysis Workspace projects are built on a data view, not directly on a connection. This means the actual analysis you run depends on the data view’s configuration. Without a proper data view, even a well-set-up connection won’t be usable for reporting or insights. The Person ID remains essential throughout, ensuring all data in a connection can be linked to individuals.

How to apply it

Start by setting up a connection to bring in datasets from different sources—like web or app events—into Customer Journey Analytics. Ensure each dataset includes a Person ID so that rows can be linked to individual users. If datasets from different channels share the same Person ID, they can be combined in one connection, enabling cross-channel analysis. This step establishes the data foundation but doesn’t define how the data will be used in analysis.

Next, create a data view on top of that connection. The data view determines which fields become dimensions or metrics and sets their behavior, such as how long they persist or how attribution is calculated. Multiple teams can use the same connection with separate data views—for example, one tailored for marketing and another for product—so each can analyze relevant data with customized settings.

Finally, when building an analysis in the Analysis Workspace, you select a data view, not the connection directly. This ensures the analysis uses the correct field definitions and attribution logic. Getting the connection and data view right separates data intake from analysis design, making it easier to manage and share insights across teams.

Mistakes to avoid

A common mistake is assuming a connection is enough to enable analysis. Without a data view, datasets remain isolated and cannot be structured for reporting or analysis. This means teams may see raw data but cannot build meaningful insights or apply attribution rules. Always create a data view to define how fields behave—such as which ones are dimensions or metrics—and to set persistence and attribution settings.

Another error is treating a data view as a one-size-fits-all solution. If different teams (like marketing and product) need different field definitions or analysis logic, they should each have their own data view on the same connection. This ensures accurate, context-specific insights. Misaligning data views leads to inconsistent reporting and misinterpretation of customer journeys. Always verify that each data view correctly defines the Person ID and aligns with the team’s analytical needs.

Quick checklist

• Confirm that the required Person ID field is defined for each dataset in the connection, as it enables person-level linking across channels.
• Verify that datasets from different channels are using the same Person ID to allow cross-channel analysis within a single connection.
• Ensure the connection includes all necessary datasets to support the intended analysis goals.
• Create a data view to define which fields will be used as dimensions and metrics, including their persistence and attribution settings.
• Assign a unique data view to each team or analysis need (e.g., marketing vs. product) so that analysis settings are properly scoped.
• Confirm that the Analysis Workspace project is built on a data view, not directly on the connection.
• Validate that the Person ID field is consistent and correctly mapped across all datasets in the connection.
• Double-check that the data view settings align with the business questions being addressed in the analysis.