Is there information anywhere on how the Shopify Plus Recommended Products API actually recommends the products? In the rendering product recommendations tutorial, it says that "Product recommendations use an algorithm that predicts and displays products that a customer would likely buy". I assume from this that the products that are recommended are based on customer profile or some sort of session cookie?
We're testing it out on our development store and noticing that there is some duplication of recommendations between products but we can't work out what that's based on. What parameters are taken into account in the recommendations?
Good question, I'd also be interested in knowing this. We just upgraded to ShopifyPlus and I wasn't aware there was a recommended products API. Thanks for filling me in :D
Since this question was first asked Shopify opened their Product Recommendations API to all stores and added some documentation pages to clarify how it works:
It states that is mainly uses two different sets of data: co-occurrent purchases (when two or more items are bought during the same act of purchase) and product description similarities.
However, this is immediately followed by quite a long list of limitations (the description based similarities won't work for anything else than english or non-Plus stores, the purchase history won't be used if you sell too many products) and if none of the optimal conditions are met, it would revert to showing random items from the same collection like the former feature that shipped with most themes.
|an hour ago|
|2 hours ago|
|2 hours ago|