Using machine learning to make real-time product recommendations


Personalized product recommendation strategies work. Multiple studies from the likes of Forrester and Gartner have found that such a strategy, if deployed effectively, can increase revenue by up to 300%, boost the conversion rate by 150% and the average order value by 50%. Little wonder that 70% of Amazon’s homepage is dedicated to product recommendations!

Fortunately, using BlueVenn’s real-time personalization tools, BlueRelevance, it’s not just retail giants like Amazon who are able to include Product Recommendations on their website. But how do they work and how can you use them yourself?

From the moment a visitor arrives on your site, BlueVenn can track every product they browse in real-time, allow the system to record information on what is selling right at that moment. This knowledge is then combined with your ‘Slot Rules’ (the rules that decide which piece from a selection of personalized content, such as banners, coupons, timers or recommendations) in order to decide what products to display.

In addition to the standard Product Recommendation option, BlueVenn enables three different recommendation types, based on a machine learning approach. The cross channel rules engine allows recommendations to follow browsers across different interaction points, targeting different customers with different recommendations.

'People Like You Buy’

People Like You Buy

In a similar way to Social Proof, where people use the action of others to shape or reinforce their own behavior, the ‘People Like You Buy’ option looks back at the history for each individual shopper and what products they have browsed. The algorithm then looks back at other people who browsed those products and recommend you the products that they end up buying.

This general purpose recommendation can be used for many different types of web pages where that you think will benefit from website personalization.

‘Frequently Bought With This’

Frequently Bought With This

This option looks at the product on the current page, then back at what people who bought this product have also bought to accompany it. For a product that has many additional accessories or complementary items, this is great to use as cross-sell recommendation tool, for use on product pages, cart checkout pages and triggered abandonment emails.

It would also work well as an additional “did you forget” page on the checkout, after the cart page.

‘People Who Viewed This Bought’

People Who Viewed This Bought

Using this option acts as a subtle endorsement of the decision to buy, by showing that others have committed to purchase. ‘People Who Viewed This Bought’ promotes the highest converting product that has been bought by the people viewing the current product. Using real-time affinity strategies like these are also a good way to show a large amount of your product inventory to browsing customers.

Different product recommendations can be aligned to different stages of the customer journey. For example, new customer acquisition can focus on special offer product recommendations, while existing customers can be targeted with recommendations based on past purchases.

Done well, recommendations will not feel like a pushy or intrusive ways to increase sales. Instead, they should feel like a natural extension of your services, helpfully tailored to their tastes to enhance their browsing and shopping experience.

Real-Time Tactical GuideThe Real-Time Marketing Tactical Guide eBook

A playbook for creating an engaging and personalized experience for your customers

This informative playbook looks at several areas of real-time marketing and personalization, including:

  • Why you need to create a dialogue with your customers to acquire data for real-time tactics
  • The benefits of getting personalization right
  • How consumers feel about real-time personalization
  • How you can use real-time marketing strategies, including triggered messaging, countdown timers, product recommendations and cart recovery 


Topics: Article