Award Confetti logo-bs logo-dxn logo-odl Magento Piggy Bank 09AEAE68-D07E-4D40-8D42-8F832C1A04EC 79C8C7E9-0D9D-48AB-B03B-2589EFEE9380 1A734D92-E752-46DD-AA03-14CE6F5DAD57 E622E2D4-3B9C-4211-8FC3-A1CE90B7DFB2 Group 19
The Breakthrough Agency.

Product Recommendations in eCommerce: best practice for Adobe Commerce & Magento

Product recommendations on Adobe Commerce or Magento giving you a headache?

As we’re a specialist eCommerce agency working on these platforms, naturally they are our area of expertise – and we have a ton of great advice to share on how to implement, maintain and improve your product recommendations on both Adobe Commerce and Magento Open Source.

Most eCommerce merchants know that product recommendations are a key tool for their stores to increase AOV and build brand affinity – and a great way to emulate one part of the ‘in-store’ experience missing in eCommerce. However… with many different tools available, as well as manual rules and AI learning, they can be a complex system to get right, with many reporting to us that systems they have in place don’t work as expected.

So to help any merchants out there struggling with this technology, we wanted to share our knowledge of the options for product recommendations on Adobe Commerce and Magento – as well as some of our best-practice advice for getting the most out of them, no matter what tools you decide to use.

Product Recommendations on Adobe Commerce and Magento: Best Practice hero image

What tools are available for product recommendations on Adobe Commerce and Magento?

Product recommendations on Magento Open Source

Magento Open Source has a number of features for product recommendations available out of the box. There are three block options, with three possible placements:

  • Related Products
  • Up-sell Products
  • Cross-sell Products

While the names may seem limiting, they are only the naming conventions, and you can actually use these blocks for any functionality you require – so there is the possibility to get creative with them.

This is a block designed to sit on the product page – essentially to showcase products that can be purchased together to increase basket value.

  • An online gardening store may want to recommend compost, pots or feed alongside plants and seeds
  • Customers purchasing bleach may wish to browse other cleaning products, like multi-surface spray or mould remover
  • A toy store may want to showcase other products in popular ranges – like Peppa Pig or Paw Patrol!

Up-sell Products

Also designed to be featured on the product page, but with a key difference – this block is designed to up-sell products that are similar, but with a higher product value.

  • A brand selling mobile phones or tablets may want to recommend the next most expensive model in the range
  • A grocery store may choose to recommend similar products from their luxury range to encourage customers to choose the higher-priced versions
  • Customers shopping for art supplies may be curious about the artist-grade versions of products while browsing the student-grade versions

Cross-sell Products

This block is designed to be featured on the basket page, acting as a final point in the journey for customers to discover additional products that could further improve their purchase experience. The types of products shown are generally similar to the Related Products block.

  • This is a great place for brands to showcase replacement consumables for products – everything from vacuum cleaner filters to coffee machine pods
  • When shopping for bed linen, customers may also want to purchase duvets, pillows or throws
  • A store selling makeup may want to recommend brushes or application tools

What are the drawbacks?

The biggest issue with the native Magento Open Source functionality of “related products” as a whole is simple – everything is manual, meaning you have to assign every single variant to every single product. For example, if you wanted to assign up-sell products and had 5 products that you wanted to link, you would need to manually assign the other 4 products to that 1 product and repeat 4 times. On a site with 100s or 1000s of products with an ever-changing inventory, this becomes unmanageable very quickly.

Product recommendations on Adobe Commerce

Adobe Commerce uses a slightly upgraded version of what is available on Magento Open source – with a feature called automated rules for product recommendations. This functionality sounds good, as it allows some of that manual logic to be removed – letting merchants set up basic rules that apply to products across the store automatically. For example, you could choose for a Related Products block to “Show any product that is within the same product type” as the product being viewed, removing a lot of manual admin processes. 

However, it doesn’t solve every issue – the rules still need to be set up, and unless you want fairly generic rules, most merchants will find they still need lots of different rules, which again can get complicated when they start contradicting each other, leaving room for error. This system may also not offer the flexibility required, especially for merchants with large product catalogues. 

Adobe Experience Cloud

Adobe have realised the flaws with the current native functionality in both systems and have created their own AI, Adobe Sensei, which powers Adobe Experience Cloud, a SaaS solution for product recommendations which is both dynamic and based on customer sales data.

Backed by the power of Adobe, this is a powerful system that is now also being used to power other features on Adobe Commerce – like their search functionality.

It is important to note that some of the third-party solutions (which we’ll detail in the next section) also use AI and machine learning along with sales data, just like Adobe Sensei, so this isn’t the only way to benefit from this technology – but for merchants who prefer to keep their tech stack within the Adobe ecosystem, this might be the best option for you.

Third-party SaaS product recommendations options for Adobe Commerce and Magento

If the native solutions don’t work for you, the alternative would be to choose a third-party system – there are several independent SaaS providers which integrate with both the Magento Open Source and Adobe Commerce platforms, so if the extra features warrant the extra monthly cost for your business, this might be the way to go.

The problem with using SaaS solutions

The biggest issue we see from working with SaaS products is that there is a heavy reliance from merchants on the automated logic – without the required logic set-up in the first place as a fall back.

Sometimes key details needed for the setup or onboarding process can be missed – either due to short deadlines, miscommunication or new staff inheriting a tech stack they haven’t had training on. We have found that some merchants are not aware that some of the key logic that makes these systems work effectively is based on sales data, and if the recommendations service is new to the website, or new products are added, it will take time for that data to be populated and thus, start working effectively.

This can cause problems for merchants, with the quality of the recommendation results being very poor for a small period of time, while that sales data required builds up.

We find that the main problems merchants run into with product recommendations are caused by two issues:

  • Logic not set up correctly
  • Lack of sales data to inform AI-based recommendations

How can you resolve this?

Despite these problems being quite common among merchants, there are some key and relatively straightforward steps that can be taken to resolve them, and ensure the best use for product recommendations when using a SaaS based solution.

Fallback logic

There is no point in having product recommendation blocks on your eCommerce site that are either not populated, or populated with a limited amount of products. There can be many reasons for this – one of the most common, particularly for new customers, is that the main logic powering the recommendations is based on a customers’ browsing history, but the customer has not browsed any products yet. 

To solve this, we would advise that merchants ensure there is always some fallback logic in place that populates the product recommendations where applicable. Testing product recommendation previews is a good way of understanding what results are expected to be returned on certain pages.

Working with your account managers

As part of your package with any SaaS product, you should be assigned an account manager that should check in regularly with you to understand how you are using their service, and what improvements can be made. We would highly suggest that you take advantage of these and chase them at least every 6 months if they are not in contact with you regularly. The sales team will know the best way to use their product, and you know the most about your products and business, so it’s a good way to move forward and find the ideal solution to any problems you might run into.

Product override

Not all services offer this – so it is worth checking – but some merchants want a hybrid solution, with most of the results coming through based on the automated rules set up, but with one or two products that should always appear in the recommendations blocks regardless. This could be just a fallback logic, but it is worth investigating as you may want certain products to appear – your biggest sellers, or a new product launch, for example.

Measure ROI

There is a cost associated with each SaaS based solution, so over time it is important to understand what financial value can be associated with your product recommendations. For example, Nosto displays the financial value associated with each individual recommendation, and you can view data including impressions, click through rates, conversion rates and total sales.

To ensure your investment in these services is financially viable, it is important to set goals for each of your recommendations and adjust the rules as needed to improve the results.  This is also a good agenda point during your regular meetings with your account manager at your product recommendations solution. 

Best practice for using product recommendations on Adobe Commerce and Magento

Where on your Adobe Commerce site should you use product recommendations?

First of all – there is no one size fits all when it comes to this! However there are several places we would advise that you always make sure to include product recommendations:

  • Homepage
  • Product page
  • Basket page

These are all areas of your site where customers are looking for products or making a purchase decision, so the perfect place to include other products they may be interested in.

Other areas you may want to include product recommendations could be:

  • Category page
  • Search results page
  • 404 page
  • Order confirmation page

Outside of the category page, which we would argue should nearly always include product recommendations, we would highly suggest adding a block to the 404 page. The 404 page specifically is usually heavily underutilised in terms of functionality and features. With the changes of product availability it’s easy to miss a redirect – or people mis-type a URL – so having some product recommendations included there is a quick win to get customers to popular products quickly, instead of just the generic ‘go back to the homepage’ link.

How can you use product recommendations on Adobe Commerce and Magento without just relying on sales data?

Depending on which recommendation blocks you use, and where you include them, to get the best results, you should aim to use a range of different logic to power each of the blocks.

Different services have different available options, but some of the most useful you could consider include:

  • Best sellers
    • Based on the sales data for other customers – showing your site’s best selling products
    • Ideal to place on the 404 page or homepage
  • Geo-targeted trending products
    • Based on popular products sold near the customer’s location
    • Can be useful for different locations, as different regions have different buying patterns
  • Landing page recommendations
    • Based on the products people viewed after visiting the same landing page on the site
    • Useful for category pages
  • Personalised Recommendations
    • Showcase the related or recommended products based on the product data stored in the system
    • Useful for product pages and cart pages
  • Browsing history
    • Based on the customer specific browsing data on the site
    • Useful for a fallback logic
  • Live feed
    • Based on what people are viewing on your site right this minute
    • Good for limited periods such as sale periods, or when a new pre-order is live, to bring further emphasis to new product launches
  • Order related recommendations
    • Suggestions based on the customers previous orders
    • Useful for homepage or category landing pages

How do your product recommendations stack up for your Adobe Commerce site?

If you’re struggling with choosing the right option for your product recommendations – or having trouble integrating your chosen service – maybe it’s time to consider changing your eCommerce agency. We support and advise our eCommerce clients on matters across their business, from partners and integrations to long-term business strategy.

To find out how we could help you, get in touch with us via or +44(0)115 7940060, or fill in the form below.