How Much are Product Recommendations Really Worth? We Did the Math

Product Recommendations: The Cornerstone of Hyper-Personalized Experiences

For most people, personalization = recommendations. The little pop-ups and carousels that put a personal spin on your catalog are what customers think about when they think about personalization. 

And with good reason. 

Recommendations strengthen engagement and drive relevancy in your customer experience. 

According to The State of Site Search in Ecommerce in 2024—a report we recently collaborated on with our partners at IMRG and Klevu—even shoppers who clicked on a product recommendation but didn’t make a purchase showed higher engagement rates, being 20% more likely to revisit the site later.  

And that’s because it shows customers that you care. 

If you can showcase products in a way that demonstrates your understanding of customers’ habits and preferences, it makes them feel valued. The more personalized you can make your recommendations—with the nifty addition of AI and machine learning to power it home—the more enamored your customers will be. 

Personalized recommendations help:

• Boost conversion rates—Relevant recommendations that confidently and seamlessly guide the customer to the product drive up conversions and grow the value of their baskets.
• Build customer loyalty—Loyalty is built on trust. If your customers trust you to recommend things they’re going to love, they’ll always come back for more. 
• Energize engagement—Finding a shiny new thing is always a bonus! The tantalizing hits of excitement that well-personalized recommendations bring are a great way to stimulate engagement. 
• Reduce cart abandonment—When you find what you need after a few clicks, there’s much less time to wait around and consider abandoning ship before you get to the cart page.

But Recommendations Are No Magic Bullet

Despite all the great things recommendations do for your customers, they can’t solve all your online retail problems. That’s because, even though product recommendations work exceptionally well for those customers who do engage with them, not all customers are inclined to engage. 

Only 13% of the 1,900 customers surveyed for our report felt that they were ‘always’ or ‘often’ influenced by product recommendations. 

That leaves a huge percentage of your site visitors either lukewarm (‘sometimes’ influenced = 34%) or completely out in the cold (‘rarely’ or ‘never’ influenced = 51%). 

So, what can you do to win over this large majority? 

Retailers must do two things. 

1. Make your recommendations even more personalized

If you can tailor all types of product recommendations to the individual needs and desires of each customer, accounting for age, location, time of day, time of year, past behavior, etc, you’ll push up your overall engagement and conversion metrics. 

2. Make your recommendations part of a broader personalization effort

Not everyone is open to having similar products recommended to them. Often, visitors just want to get to the product they came for quickly and without interruption. 

Using past behavior and other customer data across channels and interactions to  intuitively and automatically understand intent is a core part of improving the shopping experience

If you get this right, they’ll trust you, and, on subsequent visits, when they’re feeling a little more in the mood, then they’ll see a well-crafted, relevant, and timely recommendation and be much more open to engaging with it.  

Personalized Search & Recommendations: Delivering Value Together

Think about how customers interact with your site. 

While we can put them in categories like ‘Searcher’ and ‘Browser’—indeed, it’s helpful to do so to properly understand visitor intent—sometimes people don’t fall into our predefined categories as neatly as we would like. 

Real customers use your site to get what they want in several different ways, and sometimes the same customers can approach a problem from different angles within the same session. 

Research from IMRG’s personalization and AI study in 2023 showed what a customer will do once they receive product recommendations that they don’t like: Over half (55.5%) will return to the search bar to find better alternatives. 

It demonstrates the way search and product recommendation algorithms can work together and shows how an overall brilliant customer experience can give retailers a second chance. 

Use this willingness to engage to your advantage. The customer wants your site to work. When a customer clicks on a recommendation, they trust, or hope, that you understand them enough to intuitively know what it is they’re looking for, even if they don’t quite know themselves yet!

If, in general, you have an easy-to-use site with advanced search features like collaborative filtering, most customers will persevere with you even if they don’t find what they’re looking for right away. In addition to turning to the search bar, almost 27% of customers will browse product category drop-down menus if the recommendation you make for them doesn’t work out. 

But before we move on, a strong word of caution: 

Close to 25% of customers will walk if you don’t like the recommendation you put in front of them. That’s a massive chunk you’re losing to competitors. 

So what’s the conclusion from all this?  

You need to focus your energies on making recommendations personalized, relevant, and targeted and position these recommendations in the context of a broader, hyper-personalized customer experience—one in which search plays a big role. 

The Real Impact That Personalized Search & Product Recommendations Have on Your Bottom Line

And now for the math we promised you at the start: Let’s take a closer look at how much product recommendations are really worth.

In terms of cold, hard cash, the figure is $101.63. That’s the average order value (AOV) for shoppers shown irrelevant recommendations, with an average of 2.9 items purchased.

Of course, that’s before we add in all the benefits in terms of engagement, conversion, and customer loyalty that properly personalized recommendations bring. While we don’t have the exact number for that on hand, it stands to reason that by delivering personalized product recommendations based on browsing behavior and other data, retailers have the opportunity to increase average orders even more.