What are User-To-Item Product Recommendations in Betaout?

 

 

Product Recommendations are dynamic, real-time item recommendations that are generated by the users of an e-commerce store by recommendation engines.

These recommendations for ecommerce companies allow customers to find the products based on their affinity and are relevant to them quickly and easily, thereby positively impacting conversion rates.

Recommendation engines essentially leverage user behavior on an e-commerce store, to come up with item suggestions based on their browse and purchase behavior coupled with a vast number of individual attributes and product properties.

Here’s an example from e-commerce giant Amazon that you must have come across if you’re an Amazon user.

Product Recommendations

According to McKinsey, 35% of Amazon’s revenue is generated through its Recommendation Engine. Furthermore, another study by Forrester states that Amazon’s conversion rate of on-site recommendations is as high as 60% in some cases.

Given how product recommendations can boost subscriber numbers through engagement and stickiness, E-commerce stores now have the perfect opportunity to make use of the abundance of customer data available to them.

To move that data from silos, and mine it to analyze patterns, insights and actionable intelligence. For ecommerce stores looking for a robust, intelligent and highly-scalable Product Recommendation engine, Betaout can be a great solution.

In this article, we’ll dive deep into the know-hows from a technical standpoint, about one of the two types of Product Recommendations technology available in Betaout: User-To-Item Product Recommendation.

For further reading, please refer the following articles previously written on Product Recommendations:

 

What is User-To-Item Product Recommendation?

Product Recommendations essentially work on a principle of data mining called Affinity Analysis. The algorithm discovers co-occurrence relationships among activities performed by (or recorded about) specific individuals or groups.

In e-commerce, affinity analysis is used to perform market basket analysis, in which recommendation engines seek to understand the purchase behavior of customers. This information can then be used for purposes of cross-selling and upselling and enhanced product discovery.

User-To-Item Recommendations are recommendations made on the properties and attributes associated (only) with the user’s behavior and actions on your online store.

Product Recommendations

This form of product recommendations makes it easier for you to target customers across the marketing funnel, as the user’s product discovery process is based on your user’s preferences.

Here’s an example of User-To-Item product recommendation.

Product Recommendations

Types of User-To-Item Product Recommendations

User-To-Item Product Recommendations primarily work on predictive analysis and unique historical data points available for different users. Here is the logic that is followed to generate this kind of product recommendation for a user.

Product Recommendations

  • Last Viewed: For Last Viewed User-To-Item Product Recommendation type, last n products viewed (only) in the order of recency are extracted for the user. Where n is the number of recommendations you can choose to show for this attribute. Last viewed product recommendations are frequency agnostic.

  • Last Added To Cart: For Last Added To Cart User-To-Item Product Recommendation type, last n products added to cart (only) in the order of recency are extracted for the user. Where n is the number of recommendations you can choose to show for this attribute. Last added to cart product recommendations are also frequency agnostic.

  • Last Purchased: For Last Purchased User-To-Item Product Recommendations type, the last n products purchased in the order of recency are extracted for the user. Where n is the number of recommendations you can choose to show for this attribute. Last purchased product recommendations are also frequency agnostic.

  • Recently Bought: For Recently Bought User-To-Item Product Recommendation type, last 10 products that were bought in the order of recency are extracted for the user. Recently bought product recommendations are also frequency agnostic.

  • Recently Viewed: For Recently Viewed User-To-Item Product Recommendation type, last 10 items that were viewed in the order of recency are extracted for the user. Recently viewed product recommendations are also frequency agnostic.

  • User Recommendations: User Recommendations are the most enhanced and powerful of all user-to-item product recommendation types since it is hyper-personalized and tailored to the user’s affinities and preferences. Following are the attributes based on which these recommendations are generated: Location, Category Preference, Brand Affinity, Products Viewed.

 

Why use User-To-Item Product Recommendation?

If you’ve read till here, you know there essentially are two types of Product Recommendations: Item-To-Item and User-To-Item. Both of these types of recommendations have their own advantages, however, there are certain use-cases and pain-points that are only being addressed by each of these individually.

User-To-Item Product recommendations essentially allow you to leverage the data available on your customer to generate a product discovery pattern which is entirely based on how they interacted on your e-commerce store. It is more personalized, targeted, relevant and contextual from the user’s point of view as it helps them recognize their behavior on your store.

For ex: Last Viewed would remind the user of items which they were looking for, and act as a trigger for them to continue their search.

Channel Availability For Product Recommendations

Email

On-Site

Browser Push

Mobile Push

SMS

Cart Recovery

Ad Retargeting

Live Chat

Item-To-Item

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

User-To-Item

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Pages Availability For Product Recommendations

Homepage

Category Page

Product Page

Cart Page

Checkout Page

Item-To-Item

Yes

Yes

Yes

Yes

Yes

User-To-Item

Yes

Yes

Yes

Yes

Yes

 
How To Set-Up Product Recommendation Engine in Betaout?

This post will walk you through the steps required to set-up and execute User-To-Item Product Recommendation Engine for your Betaout account:

Step 1: In the Settings page, go to Recommendation Engine.

Product Recommendations

Step 2: Based on the type of User-To-Item Product Recommendation you wish to use, select the option against each type to initiate.

Product Recommendations

Step 3: Now you’ll be required to configure the type of user-to-item recommendation you have selected. Follow the steps below:

I. Number Of Products To Be Shown For This Attribute: simply refers to the number of product recommendations that you wish to be shown for the attribute.

Product Recommendations

NOTE: This step is common for all types of User-To-Item Product Recommendation you wish to activate. Kindly repeat this step for every type manually. For ex: If you wish to show more than one type of User-To-Item Product Recommendations, Kindly repeat Step 2 and Step 3 for each of the types individually.

Step 4: Click on SAVE to complete the process. Now you can successfully initiate User-To-Item product recommendations for your users.

Fallback

As the name suggests, a Fallback is a default recommendation that is fetched in place of the custom recommendation if the custom recommendation for that user isn’t available or generated for that user in the Betaout system.

When you are setting up the Product Recommendation Engine for User-To-Item type, and activate either of the 6 types of User-To-Item Product Recommendations available, in certain cases the data or a part of data would not be available for a certain recommendation that is required. For solving such a scenario, a fallback is introduced which will be shown in the place of recommendation that was to be generated by the system.

The default fallback for all User-To-Item Product Recommendations in Betaout is the top selling products on store i.e. the products which have been purchased the most number of times, overall on your store in its entirety. This would primarily be based on the frequency of purchase of the product.