Customer analytical data model

Data lies in the heart of 1:1 customer engagement, and Pega Customer Decision Hub™ requires relevant data to make high-quality decisions for them.

The customer data resides in various sources, such as a data warehouse or data lake. Typically, Customer Decision Hub requires only a portion of this data. Pega has constructed industry specific data models for customer-centric decisioning. The models are based on best practices synthesized from implementations at innovative organizations around the world, following an outcome-driven design process. This structured data model is often referred to as Customer Analytics Record (CAR) or Extended Customer Analytics Record (xCAR).

 

The main table, also known as the Spine, is a flat, structured record of the customer data in a format that is optimal for performance. Other entities, such as account data, transactional data, and action insights, enrich the main table. This structured data model provides accurate and faster business decisions and allows Customer Decision Hub to present customers with actions that are more relevant and personalized to their needs.

 

In addition to the data that comes from the data warehouse and data lake, data is also generated within Customer Decision Hub. Interaction History is the internal data layer that stores information about the interactions with customers. Customer Decision Hub uses Interaction History Summaries to aggregate the interaction history data into meaningful summaries and uses the data as predictors in Adaptive models and contact policies. Interaction History Summaries aggregate interaction data in real-time and new aggregates can be configured using Customer Profile Designer. Customer Decision Hub uses these summaries to calculate Action Insights that the Next-Best-Action Designer uses as contact policies to suppress actions.

You can also define summaries on any streamable data source that improve decision making by identifying customer intents and interests.

The data model is not a source of reporting, but rather, a source of information to help make accurate decisions. Data models should contain little to no data that is not actively used in engagement policies and predictive models. Also, as a best practice, avoid creating nested structures more than three levels deep when extending the data model. It can affect system performance negatively and complicates creating new eligibility conditions.

To implement a customer data model, use Pega-provided data model accelerators when available. You can use these accelerators during the initial setup of Customer Decision Hub, or you can download, and then apply them manually from Pega Marketplace.

During the initial phase of your implementation project, you perform a data mapping workshop.

For each industry, the data model documentation is available in XLSX format, and it includes crucial information on each entity and attribute. The workshop helps you to visualize the structure of the industry data model and to map it to your existing data model. Typically, you perform a gap analysis after the data mapping workshop to extend the data model with new attributes and entities specific to your requirements.

After you complete the data mapping workshop, run the setup wizard, and then apply the appropriate industry data model specific to your use case. Doing so generates the data model and associated rules in Customer Decision Hub. Completing the data mapping workshop early in the process gives you a head start in development and helps data architects work on database related tasks right away.

After you configure the Customer Decision Hub, create new and extend existing entities by using the output of the gap analysis.

To extend an existing entity stored in the database:

• Expose the property in the database table,
• Create the property in Pega,
• Map the property in the class definition.

To create a new entity in the database:

• Create a new table in the database,
• Use the utilities that come with the Pega platform to generate all the artifacts.
• Create a data set to access the data in the database table.

If the attribute or entity is sourced in a Cassandra dataset, you do not need to expose the property or do additional mapping in the class rule. For more information about configuring the data model, see Setting up the Customer Profile.

Use Customer Profile Designer to build the relationship among the new entities, configure your customer data model, and then ingest data from the data warehouse and data lake into Customer Decision Hub. It is also possible to import data manually by using a CSV file.

Development is a working process. As new business requirements arise, so does the need for new attributes and entities. The project team builds these new requirements in the development environment and then migrates them to higher environments through an enterprise change pipeline in a continuous development cycle.

---

This Topic is available in the following Module:

• Building the customer analytical data model v2