Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 4 Current »

The Items you may like product sets contains recommended products based on profile behavior.

Composition

Based on the behaviour of a profile, products are recommended that other profiles also viewed, added to their basket or purchased. The events used to compose the product set are:

ViewContent

AddToCart

Purchase

The following is a simplified example of how the product set for profile A is composed. To simplify the example, only the ViewContent event is used. The process below happens on a large scale with all profiles, events and products in Squeezely. In this process, factors such as time and weighting also play a role in the final result.

Input profile A in Squeezely

  • Profile A has viewed product 1

Input all profiles in Squeezely

  • Profile B has viewed products 1, 2 and 3

  • Profile C has viewed the products 1 and 3

  • Profile D has viewed products 1, 2, 3 and 4

Output

For profile A, the following Items you may like set is constructed (in order):

  • Product 3 → because the three profiles that viewed product 1 also viewed product 3

  • Product 2 → because of the three profiles that viewed product 1, two profiles also viewed product 2

  • Product 4 → because of the three profiles that viewed product 1, one profile viewed product 4

Options

The following extra options can be applied to the composition:

  • Max # of products: the maximum number of products in the set (when the maximum number is reached, the product that was first in the set goes first)

  • Exclude product set: products from another product set can be excluded for this product set

  • Exclude # last recommended: the number of recently recommended products is excluded for this product set, so that the same products are not shown constantly

  • Recency weight: the extent to which more value is attached to recently viewed products

  • Recommend based on products: the recommendations are compiled based on the product id that is sent along in the event instead of the profile's behavior

 Example recommend based on products

When you select this option, it is obligatory to send a product id in the event when composing the product set. The set will be composed based on the product id which is sent. So the relations of the product with other products is specifically looked at, instead of the total behaviour of the profile towards products.

  • Fallback set: if insufficient products meet the criteria to reach the maximum number of products, the product set can be supplemented with products from another product set.


The set can be specifically or dynamically composed with:

Product Set Filters


  • No labels