Product recommendations in Magento are deployed as a SaaS service. As you can see in the architecture diagram, the Magento product recommendation engine contains the Magento storefront and backend. The Magento storefront includes the event collector and recommendations display template, whereas the Magento backend includes the data services, SaaS export module, and the admin UI. Once the recommendation modules are installed and configured, the storefront starts collecting behavioral data. Adobe Sensei processes the collected behavioral data along with the catalog data and evaluates the associated products that are leveraged by the recommendations service.
Personalized Recommendations recipe provides suggestions tailored to an individual customer’s needs and interests. The Magento performance optimization prediction model provides insights on what products the customer’s may be interested in based on the purchase history of a customer. The eCommerce product recommendation recipe uses machine learning to analyze an individual customer’s interaction with the products in the past and create relevant and personalized product recommendations quickly and effortlessly. This creates a seamless eCommerce customer experience for the customers and also improves a customer’s overall purchasing experience, which leads to a higher engagement, stronger brand loyalty and better sales.
Adobe Sensei-led product recommendation engine uses a different combination of algorithms to showcase personalized eCommerce recommendations to an individual customer. Some of them are: content-based similarity algorithms, collaborative filtering, popularity-based algorithms, etc. The results based on these algorithms are very effective since the customers are shown the content/products that they’re really interested in, which improves their experience, leading to higher conversions and more revenue.
Types of eCommerce Product Recommendations Available
There are different types of eCommerce recommendations provided to the merchant. The admin can place the recommendations at either bottom of main content or top of main content. The bottom of main content is enabled by default and showcases the recommendations below the main content area and before any other content blocks on the page. Whereas the top of main content showcases the recommendations above the main content area just below the top navigation bar.
The following types of recommendations are provided by Magento product recommendation engine:
- Most viewed – The customer’s are recommended the items that have been most viewed by the shoppers in the span of the last seven days.
- Most purchased – The customer’s are recommended the items that are most purchased by the shoppers in the span of the last seven days.
- Most added to cart – The customer’s are recommended the items that are most frequently added to carts by the shoppers in the span of the last seven days.
- Recommended for you – The customer’s are recommended the items that are based on the individual customer’s current and previous on site behavior.
- Viewed this, viewed that – The customer’s are recommended the items that are most often viewed by shoppers who have viewed the specific item.
- Viewed this, bought that – The customer’s are recommended the items that are most often purchased by shoppers who have viewed the specific item.
- Bought this, bought that – The customer’s are recommended the items that are most often purchased by shoppers who purchased the specific item.
- More like this – The customer’s are recommended the items based on the similar content and attributes.
- Trending- The customer’s are recommended the items based on the current trends of product’s popularity.
Magento also provides backup eCommerce recommendations to fill items for which there is not sufficient input data to provide the requested recommendation. The most viewed recommendations will be shown to the customer incase if there is not sufficient input data collected for the following recommendation types:
- Recommended for you
- Bought this, bought that
- Viewed this, bought that
- Viewed this, viewed that
Default filters are defined in Magento for the most popular, trending, and recommended for you recommendation types, by using these filters Magento provides more relevant results. Admin is also provided the option to enable/disable the display out-of-stock products option, if the display out-of-stock products option is set to No, out-of-stock products will not be recommended to the customers.
In this blog, we have explored and covered different facets of the product recommendations with Adobe Sensei. The advantages that Adobe Sensei brings to business are undeniable. It is one of the powerful technologies that Adobe Commerce eCommerce platform offers to its users and equips brands to target their audience with effective strategies and achieve their commerce goals. Merchants can easily leverage product recommendations in their Magento store and fine-tune the type of recommendations they prefer for their store.
If you are looking to implement Adobe Sensei recommendations into your Magento eCommerce store, or need help in developing an ecommerce store from scratch, contact us!