Architecture and Integrations of Customer Journey Analytics

Last update: 2023-08-01
  • Created for:
  • Intermediate

In this video, find a walkthrough of the architecture of Customer Journey Analytics, including how it connects to and integrates with the Adobe Experience Platform.


Hi, welcome to this course on customer journey analytics, architecture, and integrations. In this training we’re going to talk about, the architecture of the customer journey analytics, and how it integrates with the Adobe Experience Platform. Customer journey analytics is basically analysis workspace integrated with AEP. Let’s walk through this diagram how CJA leverages the technology in AEP. Starting on the left, the data collection can come through Adobe’s SDK or other Adobe solutions. Can come from third party tools and other technologies. The data is received through streaming or batch files. And then we see here at the bottom of AEP, when the data is onboarded, it’s organized into a common set of schemas and cataloged in the Experience Data Model, or XDM. Now the XDM also controls governance on how data is used so that it complies with internal policies, contractual restrictions, and government regulations. The XDM allows AEP data lake to act like a broker of raw data when it’s requested. Here on the right CJA integrates with AEP through data connections. When CJA accesses the data lake it actually pulls a copy of the data set. It then optimizes the data to its own customization pattern. CJA uses a columnar format, which stores the data in columns, instead of rows. This allows for very fast access, filtration, and queries, and multiple sets of data from the data lake, can be curated into a single data view, and you can have as many data views as you want. Data that is onboarded into AEP is batched and uploaded every 15 minutes. When a new data connection is set up with CJA, the processing under normal loads, takes less than an hour. But for unusually high volumes of data flow, this could take up to 24 hours. From the data architecture standpoint, there are three primary tasks that need to be performed in CJA. They are, determined data needs, decide data sources, and choose identity strategy. Now because CJA data is pulled from AEP, these tasks are mostly executed in AEP. These are usually performed by the data engineer. So when we determine what data is needed for our analysis, our goal is to get a holistic view of our customer. We want the online visits, the store purchases, the customer profile, the survey results, campaign interactions, you get the idea. We want all authenticated and anonymous customer touch points. Once we determine the data needs, we then decide where the data will be sourced from and how it will be onboarded. We need to model our data to the XDM schemas by mapping each field to the schema definition. Then when we onboard the data, we know exactly what type of data we have. Then we choose how to stitch the datasets together. This is where the identity strategy comes in. A powerful feature of AEP is its ability to allow any type of data ID. So we need to determine the best way to stitch the different datasets together, by linking common IDs. The details of doing this is a whole training on its own.

Basically, the data engineer ensures that each data set has one or more person IDs identified. Then the data admin, when setting up the data connection, chooses which person ID to use for each data set. The idea is being able to say, hey, this online data has an ID that ties to the CRM data that also has a different ID, that ties to the loyalty data, that also ties to the purchase ID. The admin needs to be able to go through this process of choosing how each data set is going to be linked to each other. Once we have completed these three tasks, CJA can be set up to do our cross-channel analysis and other deep-dive analysis. Thanks for watching this training. -

For more information regarding Customer Journey Analytics, visit the documentation.

On this page