Enhance marketing intelligence with AI-integrated data
AI-fueled marketing dashboards
Take full control of all your marketing data

Guide to Effective Product Recommendation in eCommerce

A 2018 study by Anuj Kumar, a professor at the University of Florida's Warrington College of Business, revealed that product recommendations could generate about an 11% increase in sales across an eCommerce store. 

The world of eCommerce has become an extremely data-driven ecosystem. That’s why businesses leverage the enormous piles of customer data at their disposal to provide a more personalized shopping experience and ultimately maximize profit.

Product recommendation is widely used by some of the world's biggest brands, including Amazon, Netflix, Alibaba, etc. 

Netflix once revealed that its product recommendation system saves them over $1 billion every year.

According to the popular streaming service, only 20% of its user video choices come from manual search, while 80% come from automated recommendations.

Modern eCom businesses incorporate product recommendation because of the many benefits it yields, from an increase in Average Order Value (AOV) to an overall improvement in customers' shopping experience.

Indeed, improving customer shopping experience is (and should be) at the core of product recommendation. And because product recommendation effectively serves this purpose, 56% of online shoppers confirmed that they would revisit any website that offers this functionality.

This guide will explore the concept of product recommendation, covering examples and strategies you can implement to grow your eCommerce business in record time.

Let's get down to it.

Basics: What is Product Recommendation?

Product recommendation utilizes artificial intelligence to make product suggestions based on filtered data like browsing behavior, purchase history, wish lists, product views, etc.

Product recommendation example

Personalized product recommendation in eCommerce can be likened to your interaction with the shop attendants at your favorite offline store. When you ask them for a specific item, they know if it is available. But that's not all. The attendants also know other products related to what you're looking for and usually recommend the best ones.

Shopping on Amazon is another relatable example. Chances are that you've been on Amazon's website a few times to either buy a product or do research. Amazon tries to show users products in similar categories to those they'd purchased or researched and even shows personalized homepages to its site visitors.

Amazon product recommendation example

Amazon is well-known for its immersive product recommendation strategy, which it has been implementing for two decades. McKinsey & Company claims that over 35% of Amazon's customer purchases come from product recommendation.

Product recommendation is a great way to improve customer experience, boost click-through rates, increase AOV, and lock in more revenue if done right.

Types of Product Recommendation

Product recommendation comes in three categories:

  1. Content-Based Filtering
  2. Collaborative Filtering
  3. Hybrid Recommendation

Content-Based Filtering

Content-based filtering uses keywords and attributes assigned to items in a database and matches them to a user's profile for product recommendation.

The user profile is generated using data from the user's activity on the site. Such activity might include purchases, product ratings, downloads, clicks, product searches, and more.

Content-based filtering

Let's say you want to recommend a product to a visitor who just bought a laptop from your store. This laptop has a few attributes, including the manufacturer's name, a list of compatible external mouses, and more.

Also, a quick scan through the user's profile indicates that the user has previously bought a wired mouse from your store.

Now, using content-based filtering, you can recommend a compatible wireless mouse. The wireless feature is something that will probably interest the customer.

Content-based filtering is excellent because it generates highly relevant recommendations for your customers. Thus, it has a high conversion rate. Also, the data science behind this form of product recommendation is relatively easier to implement.

There are a few challenges with content-based filtering. These include lack of diversity, limited scalability, inconsistent attributes, among others.

Collaborative Filtering

To make up for the limitations associated with content-based filtering, collaborative filtering generates recommendations using data from multiple users.

Collaborative filtering vs. content-based filtering

This form of product recommendation is based on the assumption that if User A had a similar opinion with User B about Product Z, User A would be more likely to share User B's opinion on another product than they would with a random user.

For example, to generate movie recommendations for a user, the system can scan a partial list of the user's likes and dislikes and match these with several other users' tastes. Having found these overlaps, the system can then form recommendations based on what other similar users are watching.

Collaborative filtering is impressive because it requires no domain knowledge to implement and can help users discover new interests.

One of the major challenges with collaborative filtering is that it is prone to manipulation as users can sometimes rate products to rank them higher in the product recommendation system and expand the customer reach.

Hybrid Recommendation

Several experiments that compared the performance of hybrid recommendation systems with pure content-based filtering and collaborative filtering systems have shown that the hybrid system can generate more accurate product recommendations than pure systems.

Hybrid recommendations have been widely used among companies in recent times.

Netflix, for example, recommends movies to users by matching other users' watching and searching patterns (collaborative filtering.) It also offers movies with the same attributes as the ones the users had rated positively (content-based filtering.)

Hybrid recommendation

11 Popular Product Recommendation Strategies

Product recommendation springs from complex processes that mine data on users' activities within an online store to provide unique, personalized experiences for each visitor.

Every user on your website has a different need, depending on how far they've gone within your sales funnel. Identifying these needs and supplying them with suitable options is key to giving your business the boost it deserves.

Over the years, marketers have continued to populate product recommendation systems with various recommended strategies through countless trials and errors. And in this section, we will look at some of the most popular ones used by the biggest brands in the industry.

  • Automatic: This strategy is fully adaptive and recommends a product to users based on available user data, site context, the web page the user is browsing, and more.
  • Most Popular: Recommends products with the most positive interactions (views, purchases, saves, etc.) across the entire store.
  • Most Popular in Category: Recommends the most popular products within the category a user is viewing at the said time.
  • Bought Together: Recommends relevant products that are usually purchased together with the ones currently being interacted with.
  • Similar Product: Recommends products that have the most similarities with those currently being viewed or searched for.
  • Recently Purchased: Recommends products recently purchased within the category a user is interacting with. 
  • Last Purchased: Shows a list of items from a user's last purchases.
  • Purchased with Recently Purchased: Recommends a product that people typically purchase with what a user recently bought.
  • Recently Viewed: Recaps a list of products a user has currently viewed on the website.
  • Viewed with Recently Viewed Items: Showcases a list of products other similar users usually view with the items a user has recently viewed.
  • User Affinity: Recommends a product a user might like based on their preferences (browsing history, add-to-carts, wish lists, etc.)
  • Hybrid Strategy: Combines any of the above strategies to produce unique results.

8 Best Places to Implement Product Recommendations

When it comes to achieving success with product recommendation, placement plays a significant role. Your recommendation needs to be in the right place at the right time to produce the right results.

This section will look at eight places to implement your product recommendation strategies with relevant examples to inspire you.


Unless you're running traffic to a specific landing page, your homepage is usually the first point of contact for your inbound store traffic.

It is hard to craft a personalized recommendation for first-time visitors on the homepage. But one way to get around this is to use the "Most Popular" recommendation strategy or any variation of that (Trending Now, Currently Hot, etc.)

Trending products layout

For returning visitors who happen to hit the homepage, you can serve them a more personalized recommendation like "Recently Viewed," or "Recently Purchased," and "Purchased with Recently Purchased." 

Product Detail Pages

Product detail pages are where users get information about the items they're researching or trying to purchase.

The best recommendation strategies to show on these pages are "Bought Together" and "Similar Products."

Research has shown that "Similar Products" perform better than "Bought Together." So, if you can only use one, you already know which to choose.

Amazon’s similar items section

Shopping Cart Page

"Bought Together" or any of its variations works better when visitors add a product to their shopping cart. This has a positive impact on the Average Order Value as it poses an excellent opportunity for upsells.

Bought together section example


Most customers visit online stores with a specific product in mind. Thus, they make use of the search button quite often. You can suggest products that people with similar search activities have viewed.

This can offer them alternatives if they don't find what they're looking for right off the bat.

Search result page

Category Pages

Category pages store all related products in one section of the store, allowing users to access what they want more conveniently.

The best recommendation strategy for category pages would easily be "Most Popular in Category." This sparks up curiosity, encourages the user to view the recommended products and possibly make a purchase.

Most popular in the category section

404 Page

Nobody likes to land on the 404 page. Users get this when they've tried to browse a non-existent page. The best recommendation strategies could be "Recently Viewed," "Recently Purchased," or "Most Popular."

These will ensure that the customer is not left alone in the middle of a dead-end.

 404 page example


Popups can be annoying sometimes. But at crucial times, as an online store, you might need to pull it off as a last-ditch attempt.

An exit popup is a perfect example. When a user tries to click out of your website, you can implement an exit popup showing them popular discounted products or reminding them that they have items in their cart.

Pop-up example

Email Recommendation

In the event that your visitors clicked out of your website after initiating a checkout, email marketing is one of the most effective ways of bringing them back to complete what they'd started.

You can simply lead your prospects back to your shop by recommending products based on their recent shopping behavior.

Email recommendations example

5 Best Practices for Effective Production Recommendation

Product recommendation is a widely used concept in modern eCommerce. However, to get the best results out of it, you need to be creative with your strategies. Here are some pro tips to help you get the best results.

Take Advantage of Social Proof Elements

When recommending your products to your site visitors, adding social proof like reviews and confirmed purchases will give a level of credibility to your recommendations, making the users more interested in checking them out.

Amazon’s recommendations with social proof

Remove Prices from Email Product Recommendation

For prospects you're trying to recover through email marketing, it is better to focus on using high-res images, reviews, and percentage discounts to bring them back to your store. 

This is because, sometimes, prices can deter them, especially when they haven't properly gone through the product description.

Use Comparison Widgets

Comparison widgets display related items side-by-side, matching their main features against each other. This will help the users' decision-making and even make it easier for them to make a purchase.

Amazon’s comparison widget

Show Discounts and Sales

When recommending products to customers, it is usually helpful to recommend products that are currently on discounts.

When customers see this, they'd usually want to take the offer and add these items to their cart.

Incorporate Omni-Channel Marketing

Do not limit product recommendations to the bottom of your pages. Deploy them in emails, social media, push notifications, and mobile applications, among others.

A product recommendation engine, such as Dynamic Yield, will help you suggest goods to users across multiple channels. However, you’ll also need to understand what channels drive more revenue per user and how effective the recommendations are. 

A well-designed ETL system will help you streamline sales and marketing data from multiple sources and analyze the efficacy of your product recommendations. For example, Improvado can assist in lead attribution and track users’ contacts with product recommendations across all social media, email marketing tools, eCommerce platforms, and more. 

Analysts can merge data from all channels on a single dashboard to see what channels drive more leads and identify potential issues with low conversion rate recommendations. Besides, the ETL platform can help you gather customer data and feed it to the recommendation engine. Thus, it will generate quality suggestions and improve the user experience.

Improvado’s data sources for recommendation engines

This will give you more chances of getting results from your strategy.

Product Recommendation Systems

A product recommendation system is responsible for helping businesses generate personalized product recommendations using complex machine learning algorithms.

Some of the most popular ones in the industry include:

  • Clerk.io 
  • Emarsys
  • Nosto
  • Qubit
  • Retail Rocket
  • Dynamic Yield
  • Criteo
  • Adroll

Our next guide will provide an in-depth comparison between each of these systems to help you decide which one is most suitable for your eCommerce business.


With over 2 billion people shopping online in 2020, eCommerce has become an extremely data-driven industry. Innovative companies have continued to devise creative ways of using the data at their disposal to craft better shopping experiences for their customers and site visitors.

In light of that, product recommendation has proven to be one of the best ways to harness customer data for efficient marketing. We hope that this guide has shown you enough insight into the concept of product recommendation and how you can leverage it for boosting sales across your online store.

Explore more trends from the eCommerce world in this blog.

No items found.
Take full control of all your marketing data

500+ data sources under one roof to drive business growth. 👇

Book a CAll
Get up to 368% ROI

Unshackling Marketing Insights With Advanced UTM Practices


Improvado Labs: experience the latest marketing analytics technology

Be the first one to know about our latest product updates and ways they could shift workflows, performance, and effectiveness in your organization.
Track budget pacing. Our weekly ad spend is $2K per campaign. Show all campaigns that overspent or underspent this week.
Getting data from
Here's a list of campaigns not meeting your budget guidelines:
Take advantage of AI suggestions
Show total ad spend for Google Ads, Bing and LinkedIn for the last 6 months.
Our target CPL is $1,500. Show Google Ads campaigns exceeding target CPL.
Show conversions by campaign name by countries for the last 90 day
More suggestions
What would you like to ask?
No items found.
Calculate how much time your marketing team can allocate from reporting to action 👉
Your data is on the way and we’ll be processed soon by our system. Please check your email in a few minutes.
Oops! Something went wrong while submitting the form.