Personalization is one of the biggest ecommerce trends that’s quickly becoming an industry norm. In a recent survey, we found that 93% of customers expect their online shopping experience to be equal to or better than in-store — and what could be better than AI-driven product recommendations? Customer satisfaction now hinges on anticipating their wants and needs within a digital framework. Not only do they want to be treated as individuals, but they’re also expecting brands to show them content that is relevant to them quickly in their customer journey.

Using recommendation engines within your online experience can help you remain relevant across every interaction while determining your customers’ intent to increase conversion and ensure they have the experience they expect. With the help of AI, product recommendations are experiencing a renaissance.

While relevant recommendations are now expected on product pages, a lot of them still use static rules, taxonomy, or simple page view tracking mechanism. New advancement in data capture and attribution, product embeddings, and multi-device tracking, allows recommenders to be more precise, personalize and efficient. With these new tools in hand, applying the right strategy becomes paramount.

Why “brand experience gaps” might bankrupt ecommerce enterprises in 2022

Want to implement personalized product recommendations? Let’s talk recommendation strategy.

4 Product Recommendation Placement Examples

1. Recommendations on the Homepage

The homepage is a great opportunity to welcome visitors with some of your best seller products to kick off their customer journey. Shopping online feels riskier than shopping offline, especially with a new store. This is a common challenge for online stores since customers are not able to see, touch, or feel a product before purchasing it.

Take advantage of the technique known as social proof to convince customers that they are purchasing products that are appreciated by other buyers by displaying reviews, endorsements, and images of your best sellers. Knowing that a product is a best seller tends to reduce cognitive dissonance regarding buyer’s remorse.

Product recommendation examples on homepage

With the Coveo product recommendations model, you simply need to include the Popular Items: Bought (popularBought) ML parameter to your model. The model supports filtering for brands and categories in case you want to showcase best-sellers in a particular category or from a particular brand. In addition, if you display product review information, you can boost products that have a higher rating — taking advantage of the social proof principle.

2. Recommendations on Product Detail Pages

What pops into your head when you think about personalized product recommendation? Probably what you see on a product detail page. This is the most common place to find a recommended product. It’s an excellent way to showcase different yet related products from your catalog to interested users, especially those in a similar customer journey.

At this stage, customers are browsing on your site and are open to see suggestions for similar products that your product recommendation engine can showcase. You can display personalized recommendations based on similar products, products from the same brand, or even other products that are on sale that they might be interested in.

The type of recommendation you want to display on your product detail page is up to you, but make sure that the strategy behind it fits with the type of customer you have. For example, if you are an online store that sells clothing from various popular brands, you might have some customers that are loyal to certain brands. This is why you should also add recommendations based on this type of user data, not only for a similar product.

Product recommendations example on product detail pages

To implement this digital experience with Coveo, you can use the Frequently Viewed Together setting from the Product Recommendations Machine Learning model.

The following example illustrates a product listing page that contains not only one, but two different sets of recommendations. Both recommendations share the same model, but differ in the configuration.

As mentioned above, both recommendations are using the Coveo Product Recommendation model with the Frequently Viewed Together sub-model. When leveraging this strategy, you need to pass the input product SKU for the model to provide its recommendations. This will allow you to showcase similar products in their shopping experience.

3. More Suggestions on Shopping Cart Page

Depending on how you are leveraging recommendations and the number of products you have listed for sale, it is highly possible that a customer hasn’t encountered all of the products in your catalog, or the right product for them.

The cart represents the last opportunity that you have to increase the average order value by cross-selling the customer. At this stage, it is best to display products that the customer might be interested in based on what can be found in the cart.

As in the image below, if a user has a BBQ in their cart, you do not want to recommend additional barbecues, because chances are that the customer will not purchase a second barbecue. However they might be interested in some accessories — and that’s when a recommendation widget can come right in.

Product recommendation example on shopping cart page - more suggestions.

With the Coveo Product Recommendation models, you can leverage the cart recommender strategy. The model analyzes frequent buying patterns by grouping together related products that are frequently bought together in the same transaction (purchase events). When leveraging the Cart recommender strategy, you need to pass the input products SKUs in the itemIds ML query parameter for the model to provide its recommendations.

As in the image above, if a user has a BBQ in their cart, the model is not recommending additional barbecues, because chances are that other customers didn’t purchase a second barbecue. This is why we are seeing BBQ accessories.

4. Recommendations on No Results Page

Do not miss the opportunity to recommend products when a user searches for a product that does not exist on your site. A visitor might not find the product they had in mind, but this does not mean that you can’t suggest something to them with your recommender system. Consider featuring popular products from your store. Using user data, the right recommender system can even drive you to recommend products that the user would like, based on similar user profiles – this is collaborative filtering

If you are a Coveo customer, you can achieve this with the Popular Items: Viewed submodel. No additional ML query parameters need to be configured for the model to provide its recommendations.

You might want to check out a previous blog post that we have about this: Avoid No Results Pages, Display Popular Results Instead.


Now that you have seen some strategies, all that is left is to choose what fits best for you site.

Coveo has a host of recommendation strategies ready for you to choose from that fits your business purpose.

Want to start creating more personalized recommendations? Learn more about how Coveo delivers relevance in commerce.

Dig Deeper

Learn more about how to recommend products to first-time shoppers in this 6-minute read.

Predict what people need, before they even know they need it—with Cove’s AI-powered recommender.

Learn more about all the different types of recommendations you can embed throughout your ecommerce experience in our ebook, Ecommerce Alchemy: The Periodic Table of Recommendations.

Introducing Ecommerce Alchemy: The Recommendation Periodic Table

Have a minute? We’d love to get your feedback.