What Is Adobe Customer Journey Analytics — and Why It Changes How You Measure Marketing

If you have spent any time in Adobe Analytics, you know the ritual. Someone asks why campaign numbers from the email platform do not match the numbers in the reporting dashboard. You pull the data. You look at the data. You send three Slack messages. Forty minutes later, you are still not sure which number is right.

Adobe Customer Journey Analytics (usually shortened to CJA) was built to address a specific version of that problem. Not the politics of it, not the tooling sprawl, but the underlying data architecture issue that makes cross-channel measurement so consistently unreliable. This article is a plain-language breakdown of what CJA is, how it differs from traditional Adobe Analytics, and why the foundation it sits on matters more than most people realize.

The problem with the old model

Traditional Adobe Analytics was built around sessions. A visit starts, events happen, and a visit ends. That model works fine when your customers live inside a single browser on a single device. In practice, they do not. A customer reads your email on their phone during their commute. They do not click on anything. Three days later, they search for your brand on a laptop, browse your site, add something to their cart, and abandon it. A week after that, they see a retargeting ad on a tablet and convert.

In a session-based model, those three interactions are three different people. The email gets no credit. The search might get it all. Your marketing mix looks nothing like reality. This is not a new problem. It is just a problem that has become harder to ignore as channels have multiplied and customer journeys have stretched over weeks rather than minutes.

What CJA actually is

Adobe Customer Journey Analytics is an analytics solution built on Adobe Experience Platform. Instead of organizing data by sessions and page hits, it organizes data by people, and it stitches together behavioral data from any source (online, offline, call center, loyalty programs, point of sale) into a single, queryable timeline per customer.

The shift in technical architecture is worth understanding, even at a high level. In traditional Adobe Analytics, data lives in report suites — relatively siloed containers that made cross-property analysis difficult. In CJA, data lives in Adobe Experience Platform as datasets, structured around a standard schema called XDM (Experience Data Model). Connections in CJA pull from those datasets. Data Views define how that data is presented to analysts.

A few changes in terminology signal how fundamental the rethink is. What used to be called Classifications are now Lookup datasets. Customer Attributes are now Profile datasets. Hit, Visit, and Visitor containers have been replaced by Event, Session, and Person. These are not just rebrands. They reflect a different underlying philosophy about what a unit of measurement should be. CJA does not ask “what happened in this session?” It asks “what has this person done across every touchpoint, in what order, over what time period?”

That shift opens up analytical questions that were practically impossible before: What was the full journey of customers who converted in Q3? How do offline purchases affect online behavior in the 30 days that follow? Which marketing channels tend to appear earliest in journeys that eventually convert, versus channels that appear at the end?

The data architecture underneath it

CJA runs on Adobe Experience Platform, which means it inherits AEP’s approach to data ingestion, identity resolution, and governance. That is worth understanding because it affects what you can do with CJA — and how clean your data needs to be before it gets there.
Data flows into CJA through Connections. Each Connection pulls from one or more datasets in AEP. Those datasets can come from Adobe Analytics via the Analytics Source Connector, from first-party data streams via the Adobe Web SDK, from third-party sources via batch ingestion, or from data warehouses like Snowflake, BigQuery, and Azure Databricks (with data mirror functionality now in development).

One of the more consequential changes for marketing teams is how classifications work. In legacy Adobe Analytics, classifications were notoriously fragile — rule-based lookups that transformed tracking codes into readable campaign metadata. In CJA, that functionality lives in Lookup datasets, which are joined against your event data at query time. This is structurally cleaner but depends on the Lookup datasets actually containing accurate, complete, and consistently structured data.

This is the part most CJA migration guides skip over. The architecture is more flexible. That flexibility amplifies whatever data quality problems you already have.

AI is now part of the product

CJA has added a set of AI-powered capabilities that have moved from optional add-ons to core parts of the product. The Data Insights Agent lets analysts ask natural-language questions and get visualizations without having to build reports manually. Intelligent Captions generate automated plain-language summaries of trends and anomalies in Analysis Workspace. Anomaly Detection flags unexpected shifts in metrics in real time.

More recently, the Data Insights Agent was integrated with Microsoft Copilot, which means CJA data can now be queried directly inside Teams, PowerPoint, and other Microsoft 365 tools. For organizations where business stakeholders live in Microsoft’s ecosystem, that is a meaningful change in who can access analytics without routing requests through an analyst.

Adobe has also been adding Guided Analysis capabilities (structured analytical templates for specific question types, such as retention, funnel analysis, and trends) that make CJA more accessible to non-technical users. The pattern is consistent: the platform is being built to put analytical insight closer to the people who need to act on it, with less friction in between.

The part most organizations underestimate

Here is where most CJA conversations focus on the wrong thing. Teams spend months on the migration architecture — schema design, identity stitching, data view configuration — and relatively little time on whether the campaign data flowing in is sufficiently consistent to produce reliable analysis.

CJA’s cross-channel model is powerful because it joins data from many sources. But that joining process depends on shared identifiers and consistent metadata. If your campaign naming conventions vary by channel, region, or team (e.g., if your UTM parameters are filled in inconsistently or your tracking codes follow five different taxonomies depending on which agency or business unit created the campaign), CJA will faithfully stitch all that noise and nonsense into your analysis.

The AI features do not fix this. The Data Insights Agent will generate a response to your question about campaign performance. That response will reflect whatever is actually in the data. If the data is fragmented, the insight is fragmented, and it will look clean and confident while being wrong. This is not exactly a CJA problem. It is a data standards problem that CJA makes harder to ignore. The more sophisticated your analytics platform becomes, the more consequential upstream data quality gets.

That is the context for why how data gets into CJA matters just as much as what CJA can do with it once it arrives.

Up next:

Part 2 of this series covers what it looks like when campaign metadata is standardized before it reaches CJA — and how Claravine’s new integration with Adobe Customer Journey Analytics closes the gap between campaign creation and clean, reportable data.