Better understanding

ABTasty recommendations are made up of default algorithms that are trained using data from your integrations (analytics, catalogue, CMS, etc.).


When a recommendation banner is displayed on a website, the website code calls the ABTasty Recos servers with input parameters (such as the article displayed on the page, the articles seen by the visitor, etc.) to retrieve the relevant list of articles in the given context.


This call applies the algorithm contained in the recommendation to the data provided.


For example, you could create a recommendation banner (recoID X24) which displays similar items on a product page (TSHIRT1) filtered on the items the visitor has already seen (TSHIRT24, TSHIRT49).


The page code will then call the Recos server with the following entries:
reco = X24
viewed_item = TSHIRT1 (the SKU of the product currently viewed)
viewed_items = [TSHIRT24, TSHIRT49] (SKUs of previously viewed products)
The server will retrieve the configuration of reco X24 and call the algorithm it contains, ‘Similar items’ to product TSHIRT1, then filter the results of this algorithm to subtract TSHIRT24, TSHIRT49.

 

Was this article helpful?

/