Machine Learning Content Models¶

These content models output details are constructed from CDP’s machine learning models. The content models are:

• Machine Learning Segments: For 1:1 machine learning models where each customer receives only one score per model. For example, Likelihood models such as likelihood to convert, engage, buy, and Clustering models such as behavior and product clusters. This content model outputs details about the ML segment, which can be a decile score for likelihood models, or a descriptive cluster name like ‘Discounters’ or ‘Swimwear’.
• User to Product Recommendations: Outputs a list of all products recommended to the customer by CDP.

Product-Oriented Content Models¶

These content models output details about products linked to customers. 

The content models are:

• Last Products Purchased: Outputs a list of all products purchased by the customer on the last transaction date only.

  For example, Customer ID 12345 makes the following purchases on the given dates:

  Transaction Date	Product 1	Product 2	Product 3
  January 1st	A	B	C
  January 10th	M	N	N/A
  January 15th	X	Y	N/A

  A CDP user may use the model expecting to capture all products from all the three transactions as depicted in the preceding table. However, the model captures the products purchased on the last transaction date only, that is, January 15th.

  If you require the model to capture products from all the transactions, contact Support to enable the transaction details content model.

• Last Products Browsed: Outputs a list of all products browsed by the customer.
• Last Abandoned Products: Outputs a list of all products that were added to the cart by the customer after their most recent transaction.
• User to Product Recommendations: Outputs a list of all products recommended to the customer by CDP.
• Favorite Products: Outputs a list of top N most commonly purchased products, ranked by quantity and followed by revenue. The default value of N is 10. You can update the value in the CDP user interface. These products are purchased two or more times in the past X months or days. Here, X is the timeframe that you can select.
• Subscription: Outputs the top N subscriptions and shares the user-defined field values for subscription, product, and subscription-product mapping attributes. The default value of N is 10. You can update the value.

Organization-Oriented Content Models¶

These content models output details about organizations or stores linked to the customers. 

The content models are :

• Closest Store Details: Outputs the physical organization or retail store that is located closest to the customer geographically (calculated by comparing the center of the store’s zip code and the customer’s primary addresses zip code).
• Primary Store Details: Outputs the physical organization or retail store that the customer has transacted at most often (ordered by transaction count).

Customer-Oriented Content Models¶

These content models output details of abstractions above or below the master customer. For example, a household is above a master customer since a household encompasses multiple master customers. Child customers are below a master customer since multiple child customers or source customer records are de-duplicated by CDP to construct a master customer. Those content models are :

• Household Details: For customers that have householding turned on, this outputs the household that the customer belongs to.
• Cross-Brand Customer Details: For customers that utilize master customer groups, this outputs details about the master customer group that the customer belongs to.
• Child Customer Details: Outputs a list of all child customers that were de-duplicated to construct the master customer.

Calculated Content Models¶

These content models are calculated at the time of campaign execution and allow for more dynamic content options. The content models are :

• Campaign Details: Outputs details about the CDP campaign for every output record.
• Echo: Outputs the exact same text (string) for every output record. The user inputs a header name for the column and some text that is repeated (echoed) for every output record.
• Event count in a time frame - Browse Product/Product Category: Outputs the total number of event occurrences in the refinement conditions and a specific time frame.
• Event count in a time frame - Email Actions: Outputs the total number of event occurrences, such as emailClick, emailOpen, and emailSent, in the refinement conditions and a specific time frame.
• Revenue in a time frame: Outputs the total amount spent by a customer in the refinement conditions and a specific time frame. You can use refinements like Line Type and Line Subtype. The possible value for Line Type is Purchase. Possible values for Line Subtype are Demand, Canceled, Shipped, and Returned.
• Transaction count in a time frame: Outputs the total number of orders by a customer in the refinement conditions and a specific time frame. You can use refinements like Line Type and Line Subtype. The possible value for Line Type is Purchase. Possible values for Line Subtype are Demand, Canceled, Shipped, and Returned.

Content Models for Promotions¶

For customs that utilize CDP’s coupon distribution features and coupon history, these are the content models that allow you to include coupons in your output content:

• Distribute new coupons: Link or assign new and previously undistributed coupons to your audience from the available coupons data store, then distribute them to your audience.
• Redistribute existing coupons: Distribute coupons that are already assigned or linked to individuals and stored within CDP’s coupon history. You can use this to remind individuals about previously communicated but still unexpired and unredeemed coupons, or distribute loyalty rewards for the first time.