Most organizations that moved to CJA moved the problem with them.
Customer Journey Analytics is a genuinely more capable platform than Adobe Analytics. Whether it returns on that capability depends almost entirely on what happened before and during the migration: whether the XDM schema was designed from the organization's data requirements, whether Person ID strategy was decided before Connection configuration, and whether Data View governance was built to hold.
Three states — same underlying architecture problem.
Organizations in the Adobe analytics stack arrive at the same set of architecture decisions from different points in the journey. The decisions are identical regardless of where you're starting from.
Planning the move to CJA and want the architecture designed before implementation starts
The XDM schema, Person ID strategy, and Data View governance need to be decided before a line of Web SDK code is written. Organizations that skip this design phase spend the first year in CJA correcting architecture under production pressure.
In CJA but running Adobe Analytics in parallel, and the CJA numbers don't match AA yet
The dual-running state has a cost: two implementations, two data models, a team that has to maintain both. The faster it resolves, the lower the cost. Resolution requires the architecture decisions the migration started without.
Landed in CJA and the platform isn't returning what it was bought to deliver
This is the most common state. CJA is live. The Data Insights Agent is available. The schema inherited from the AA migration doesn't support the analysis the platform is capable of. And nobody wants to rebuild it now.
CJA is more capable than Adobe Analytics. The architecture that makes that capability available is what most migrations didn't build.
A well-architected CJA implementation can answer questions that were structurally impossible in Adobe Analytics — cross-channel journey analysis, person-level attribution, direct integration of offline and CRM data.
The most common post-migration state: CJA is live, the Source Connector pulled in AA historical data, the Data Views mirror the old AA report suites, and the team is waiting for the cross-channel insight the platform was bought to produce. That analysis isn't available — because the schema doesn't support it.
XDM schema design determines every dimension and metric available. Person ID strategy determines how journeys stitch across devices. Data View governance determines attribution models and reporting alignment.
These three decisions dictate CJA's analytical ceiling. And most organizations make them by default during migration.
Every layer of the CJA stack has an architecture dependency in the layer beneath it.
The six layers of the Adobe measurement stack each have specific architecture decisions that determine whether the layer above them returns what it's capable of.
Deploys Web SDK and manages data elements, rules, and publishing. Configuration determines what data the Web SDK collects and how it's structured before hitting the Datastream.
Adobe's replacement for AppMeasurement. Streams XDM-native data via Datastreams to AEP. The critical difference: Web SDK produces data in the organization's XDM schema, not in legacy schema.
The data model that structures everything in AEP. Every dimension and metric in CJA Data Views traces to an XDM field. Fields cannot be removed once data flows into them.
The identifier CJA uses to stitch events into a coherent customer journey across devices, channels, and time. Determines what cross-channel analysis is possible.
The Connection links AEP datasets to CJA. The Data View is the reporting governance layer — controlling dimensions, metrics, attribution, persistence, and session definitions.
CJA's analysis environment and AI features (Data Insights Agent, Intelligent Captions). All AI features query the Data View — output quality is determined entirely by the architecture beneath.
The architecture gaps that show up as analysis problems.
These are the consistent patterns across Adobe Analytics and CJA environments — the decisions that weren't made explicitly, and that appear as data quality or reporting problems months or years later.
The Data Insights Agent is a force multiplier. What it multiplies depends on the Data View.
Every AI feature in CJA queries the Data View. If the underlying data has schema gaps, identity fragmentation, or unresolved attribution mismatches, those issues appear in the AI output as confidently as accurate insights do. The AI layer doesn't fix architecture gaps upstream of it.
Natural-language queries, visualization generation, anomaly exploration. GA June 2025. Output quality is determined by Data View governance and XDM schema design.
Four entry points — one underlying architecture problem.
The right starting point depends on where the gaps are showing up. All four paths begin with the Assessment, which maps the current state of the environment.
XDM schema, Person ID strategy, and Data View governance — designed before implementation starts. The Assessment maps the current AA environment against CJA requirements.
- Implementation team has clear blueprints
- Migration takes one pass instead of two
If CJA isn't returning what it was bought to deliver — the architecture decisions are where to look first.
The Measurement Architecture Assessment maps the current state of the Adobe environment: XDM schema design, Person ID strategy, Data View governance, Web SDK implementation status, and the gap to be closed.