What are "Scenarios" and how do they impact the recommendations?

Beam Scenario’s is a collection of business rules associated with a specific recommendation type such as User to Items.

In other words, a Scenario can contain all the following information regarding how your recommendations will be computed and delivered to your customers.

• Reranking OR Diversity: Reranking rules reorder the recommended items according to their property values.
• More Info: Reranking on Item Property.
• Filters: Filters remove the items satisfying certain conditions from the recommendations.
• More Info:
• Filtering on Item Property — Crossing Minds API v1.0.0 documentation .
• Exclude Rated Items: this filter determines if rated items are to be excluded from the returned items.
• Algorithms: the models used by the API to generate the recommendation you are evaluating or deploying
• More Info: http://b2b-api-documentation-staging.s3-website-us-west-2.amazonaws.com/endpoints/model.html?highlight=algorithms

What’s the Impact?

In short, instead of calling the recommendation rules individually (query filters, reranking, diversity, and other important business rules), you can save these configurations under a single name. Making it easy to make one call to the API using the Scenario name to apply the recommendation configuration  VS having to update the front end API call with the new configurations in the payload after each update is made to the model.

For instance, let’s say the goal is to increase the AOV (average order value) and so the product team wants to create specific rules for the recommendations that they intend to display during the checkout phase of they site. To achieve this they would like to always have filters on the prices, reduce the diversity around what type’s of products to recommend or maybe the Crossing Minds team has created a custom algorithm to boost your sales at this precise moment of the customer journey.

Through the dashboard, you can create a Scenario called “Check Out AOV” that includes the specific business rules, which can then be used to customize the user checkout flow.

Once saved, the new logic is applied without requiring any dev help because the call itself remains static while the logic is updated on the backend giving the team more flexibility to make changes to match business goals.

EX of API Call:

• "curl https://api.crossingminds.com/recommendation/users/1234/items/scenario="Checkout"

VS

• "curl https://api.crossingminds.com/recommendation/users/1234/items/amt=10&filters=price:lt:10&filters=genres:eq:drama&filters=tags:in:family,fiction&filters=poster:notempty&exclude_rated_items=true&algo=upsell"

It is important to note that Scenarios can not be passed from one recommendation type to another. For Example: The Scenariosimilar items can not be used for user based recommendations.

Multiple options can be selected.

‍