Types of product recommendations to personalize the customer experience

Types of product recommendations to personalize the customer experience Headshot

By Simon Nooranvary, Customer Success Consultant

Product recommendations are one of the best ways for digital marketers to create a personal customer experience online.

It can be overwhelming to sort through the products available on even modest eCommerce sites. Filters and searches are useful, but require effort from the shopper to enter the right terms and scan the results. Product recommendations provide a more curated experience based on wisdom of the crowd, customer data and business rules.

The benefits of well-implemented product recommendations are clear. Customers more easily find what they’re looking for, reducing the frustration of the shopping experience. And shoppers who enjoy a seamless, relevant experience are more likely to convert and stay loyal to your brand.

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

A successful product recommendations strategy combines three elements:

  • Collecting the right customer and product data to show shoppers the most relevant products.
  • Delivering the most suitable recommendation type for the shopper’s stage in the purchase journey.
  • Fine-tuning recommendations to meet your business needs.

Here, we’ll walk through the different recommendation types available, why they’re effective, and when to use them.


1. "People like you buy"

This tactic looks at products that the shopper has bought, and compares with other shoppers who've viewed those products, then uses a machine learning algorithm to recommend the most likely eventual purchases.

Why does it work?

It’s based on the fact that people who’ve had similar preferences in the past will tend to like similar things in the future. These recommendations derive from actual behavior and don’t rely on machines understanding the exact nature of each product.

By including a header such as ‘People like you buy…’, you also appeal to consumers’ desire to follow the wisdom of the crowd and feel part of a tribe.

This is a great general-purpose recommendation type for use on many different types of web pages from the homepage to the product detail page, and in emails. It helps customers at the research stage of the buying journey, where they might not know exactly what they’re looking for.

Recommendations can be used to increase engagement with booking abandonment emails, like this one from Cottages.com:

Personalized product recommendations in a cart abandonment email


2. "Frequently bought with this"/"Purchased together"

This looks at the product on the current page. It then looks back at what people who bought this product have ended up buying with it.

Why does it work?

Customers are reminded of accessories and complementary items that they might have forgotten when they filled their cart. This is an ideal recommendation to increase order value at the checkout, after the cart page.

It’s also a great way to add value to post-purchase emails, by showing shoppers the products that they are likely to want and need directly after their original purchase.

Amazon is known to use this kind of recommendation very effectively to increase basket value:


Frequently bought together product recommendation example


3. "After viewing this, people buy"

This recommends products that shoppers often buy after they view the product on the current page. A machine learning algorithm recommends the products that shoppers who viewed this product most often went on to buy.

Why does it work?

Similarly to “People like you buy…”, this type of recommendation leverages the fact that people who share a preference for one product are likely to agree on other products too. It anticipates the products that are most likely to lead to a conversion.

This lets you show a larger amount of your product inventory to web browsers. Customers get to see popular items that they might not have thought about when they made their initial search.

Here’s an example from a Base London product detail page, showing customers the products most likely to lead to a purchase:


People went on to purchase product recommendation example


4. Social proof recommendations

This recommends products that are trending with other shoppers. These could be your bestsellers, most browsed or most frequently carted items at the moment.

You can add popularity messaging to increase urgency and reinforce the idea that these items are generating a buzz with fellow shoppers. Here’s an example from Victorian Plumbing using trending icons:


Trending products feed with popularity and urgency messaging example


You can add another dimension of social proof by including star ratings in product suggestions. This email from FFX tools recommends the best-selling batteries, personalized to reflect the shopper’s favorite brand. Customer ratings are used to build trust and encourage click throughs.

Example of trending product feed with rating stars in email


Why does it work?

These suggestions can work well for customers whose preferences you don’t yet know. Since these products are popular with existing shoppers, there’s a fair chance that they will appeal to new customers too. Consider adding trending product feeds to your homepage and newsletters.

Labeling products as ‘trending’ or ‘most popular’ boosts the power of social proof recommendations. When making a purchase decision, consumers prefer to follow the wisdom of the crowd and choose similar products to their peers.


5. Similar products

This recommends products which appear similar to the product on the current page, based on the contents of product details - including name, description and tags.

Why does it work?

This type of recommendation is ideal for customers at the research stage of their journey. Shoppers see a range of alternative products similar to the one they’re browsing, so they can find the one that best suits their needs.

It’s also a perfect option for retailers who frequently add new or one-off products to their inventory: you can immediately include recently added items in your recommendation strategy, without any manual tagging or customer behavioral data.

Other applications include online publishing and jobs boards, where browsers want to see more articles or listings like the one they’re viewing. Here’s an example of similar job suggestions from recruitment website SecsintheCity:

Similar content recommendations recruitment site example

6. New arrivals

These are the latest products to be made available – whether that’s across the store, in the browsed category, or from the customer’s favorite category.

Why does it work?

There’s a reason consumers flock to buy the latest iPhone every few years. We have an in-built expectation that something new must be better that what came before. Recommending your newest products also builds excitement around the shopping experience.

This is especially true in sectors driven by fast-moving trends, such as fashion and beauty. 

Latest product recommendations work well on the homepage, category pages and in bulk marketing emails. For more impact, suggestions can be filtered to show new arrivals in the shopper’s preferred category.

Here’s an email from Harvey Nichols, inspiring recipients with the latest beauty products:

New product latest arrivals email newsletter example

7. Related products

Not to be confused with similar products, this recommendation shows products that you have defined as related to the current product for merchandising purposes.

Why does it work?

This is great for cross-sell and up-sell where the merchandiser needs full control over what products appear together. For instance, it can be used to inspire shoppers with items that complement the one they are browsing.

Fashion giant ASOS is known to use this tactic to great effect on product pages:


Related product recommendations for merchandising on product detail pages


8. Personalized - frequently browsed products

Recommends products that this individual shopper has browsed frequently.

Why does it work?

This type of recommendation is particularly effective because it reminds customers about products that they are already interested in, but haven’t carted yet.

This is a great way to engage busy shoppers who might have been distracted before carting their favorite product. It also harnesses the power of familiarity – customers tend to prefer products they have seen multiple times.

You can use this recommendation type on the homepage and in personalized marketing emails to target frequent browsers who haven’t yet made a purchase.

This is a great option for newsletters where it’s appropriate to show a mix of different product types. Here’s a newsletter from Bear & Bear featuring personalized product recommednations:


Personalized product recommendations in ecommerce email newsletter


Get more out of product recommendations

Advances in machine learning mean it’s easier than ever to serve the most relevant products to customers at the moment when they’re most engaged. AI lets marketers automate eCommerce personalization at scale – so you can devote more resources to producing engaging content and making smart business decisions.

But there are limits to the brilliance of AI, and sometimes a human hand is needed to guide the technology. Imagine these situations:

  • You’re a retailer selling multiple brands, and need strict control over which labels appear side-by-side.
  • You have an excess stock of one particular product that you need to clear before next season.
  • You want to boost the revenue from product recommendations by only showing your highest-margin products.

The most effective product recommendation engines combine the power of machine learning with the ability to filter recommendations based on your priorities. The machine takes on the difficult task of picking out the most engaging products, while marketers keep full control over how content is displayed.


First step to personalization

Experimenting with a mix of personalized and crowdsourced recommendations, social proof and your own business rules, you can achieve a level of personalization that was once only possible in physical stores.

And product recommendations aren’t difficult to implement. The right solution will work with your eCommerce platform, customer data platform and ESP to help you integrate personalized product suggestions across your online channels.

For more tips on getting started, download the dedicated ebook:

Download guide to product recommendations
Types of product recommendations to personalize the customer experience Headshot

By Simon Nooranvary

Customer Success Consultant

As Customer Success Consultant at Fresh Relevance, Simon ensures that clients are achieving their maximum potential from the real-time marketing platform, and helps them overcome any technical difficulties.