FROM CACHE - en_header

Is the Shopify product recommendation algorithm somehow extendable/customizable?

New Member
3 0 0

Hey there,

We are planning to sell Japanese alcohol ("Sake") online and we are looking for product recommendation algorithms.

Looking at the Shopify product recommender there are some limitations which make it not suitable in the long run:

  • No consideration of extra product meta data fields (e.g. taste profile, aroma etc.)
  • No consideration of extra customer meta data fields (e.g. personal taste profile)
  • No consideration of product ratings/comments after purchase

Are there any possibilities to extend the Shopify algorithm (e.g. feeding more data points into it)?

Other Shopify apps for product recommendations seem to have the same issues as above.

Thanks so much in advance.

Kurashu (


Replies 4 (4)
5 0 3

Hello kurashuJP,

There are mainly 2 kinds of recommenders (see here for more details:

- Content-based recommenders try to find recommendations based on item's features (for instance, if I am visiting a blue product, a content-based recommender will promote other blue products).

- Collaborative filtering recommenders try to find recommendations based on the visitor's customer journey (for instance, if a product is often added to cart or converted or clicked or visited, it will be promoted against other products).

Each strategy has its pros and cons: collaborative filtering for example is not ideal for cold start issue because it needs to gather any sufficient data to propose relevant recommendations. Content-based recommenders do not have this problem but can be more static and not take into account products that visitors really appreciate.

From my understanding, Shopify is using content-based recommendations (it enables them to avoid the cold start problem) but indeed it is not suitable when product ratings have to be included in the equation. In this case, your requirement is more on collaborative filtering and, to do that, you will have to considerate specialized third-party solutions. I can advise you on various possible solutions that best suit your needs (to be honest, I work for a digital company that proposes these features).



New Member
3 0 0

Hi Florian,

Thanks for your reply.

Indeed Shopify is using content-based algorithms. I guess also it is not possible to feed in extra data points into the Shopify algorithm.

At first I guess it could make sense to go with the Shopify algorithm and wait for sales & customer interactions to pick up. 
When enough data points are available, I think it would make sense to shift to a proprietary system or 3rd party service. 

What is the service name to consider?



5 0 3
Hi kurashuJP,

Thanks for your reply.

I was maybe a bit unclear in my last ticket: collaborative filtering do not
necessary need only customer interactions. It is based on real-world
interactions: visited pages, items added to cart, abandoned carts, etc.
That is the reason you must setup a script on your site that tracks these
interactions and sends them to a third-party provider. This third-party
service will analyse this data and propose recommendations based on it.
That is also why it requires a few days before being completely relevant.
The sooner you setup such a script, the more it will pay off for you.

I can give you the service name to consider on a private message. Can you
activate it (

Best regards
New Member
1 0 0

Hey ,

Use AI Personalization, Buyer Persona, Upsell Recommendation Bands - " ShopSense – Buyer Persona " App

 is an AI-powered Buyer Persona and User engagement platform that empowers marketers and sellers to enable the most human intuitive discovery of products. Understand your buyers’ preferences and relationships with your products to enable hyper-personalization through their shopping experience.

Following features possible

- Automatic product tags widget
- Automated Product Discovery and search
- personalized recommendations with upsell bands and bundle products