Setting Up Recommendation Scenarios, Recommendation Models and Model Types

Objective

After completing this lesson, you will be able to successfully set up recommendation scenarios and models to deliver recommendations to consumers and evaluate its effectiveness using the pre-delivered key figures.

Recommendation Scenarios

A recommendation scenario enables host systems to post interactions and receive recommendations.​

You can use Recommendation Scenarios to view and edit the settings that enable host systems to post interactions and receive recommendations. The following standard delivery recommendation scenarios are available:​

  • Top Sellers in Web Home Page
  • Most Viewed Items in Web Home Page
  • Top Sellers for E-Mail Campaigns
  • Most Viewed Items for E-Mail Campaigns
  • Cross-selling in Web Product Details Page
  • Cross-selling in Web Checkout Page
  • Items Viewed Together in Web Product Details Page
  • Factsheet recommendations
    • Account Buying History
    • Browsing Favorites

For more information, please see the documentation about Implementing Recommendation Scenarios

Creating Recommendation Scenarios

Screenshot of Recommendation Scenarios application.

Recommendations (SAP Business Technology Platform)

When you create a new recommendation scenario, you will have to select the communication medium and the Model type.

If you want to use recommendations in external systems like mobile apps, the Recommendation scenario will contain the Security detail like recommendation scenario URL and Security string.

Public OData API (API_MKT_RECOMMENDATION_SRV) that allows a client system to obtain recommendations from the SAP Marketing Cloud using the SAP Business Technology Platform.

The following link will provide more detailed information on how to implement recommendations using the BTP platform.

Recommendation Scenarios KPIs

You can evaluate the effectiveness a recommendation scenario using the key figures of impressions, click-through-rate, and conversion rate. These key figures are generated for the active model as of the model activation date.

Screenshot of Recommendation Scenarios application, with a box highlighting the Impressions, Click-Through Rate and Conversion Rate columns.

Read more about the key figures below:

  • Impressions: The number of times a scenario has provided recommendations to a consuming application. A single impression is generated each time the consuming application fetches recommendations.
  • Click-Through Rate: The proportion (percentage) of recommended items that are clicked on by consumers.
  • Conversion Rate: The proportion (percentage) of recommended items that are purchased by users. This is calculated by dividing the number of conversions by the total number of click-throughs. When a model expires and a new model becomes active, the values associated to each key figure are reset.

Note

When a model expires, and a new model becomes active, the values associated to each key figure are reset.

There are no key figures available for offer recommendations. Models that use a model type with SAP Marketing Offer Content defined as the Recommendation Type don’t generate any key figures.

Recommendation Models and Model Types

Watch the video below to learn more about working with recommendation models:

The recommendation models have different activation and generation statuses.

To learn more, please check out this documentationon the Help portal.

Algorithm Step Types

To learn about the available algorithm step types, watch the following video:

For additional information on how to resolve a Failed model generation, check out this SAP Note.

Log in to track your progress & complete quizzes

Learning

SAP FacebookSAP YoutubeSAP LinkedIn