Product Recommendations

Product recommendations are a powerful marketing tool you can use to increase conversions, boost revenue, and stimulate shopper engagement. Product recommendations are surfaced on the storefront in the form of units such as “Customers who viewed this product also viewed”. Adobe Commerce product recommendations are powered by Adobe Sensei, which uses artificial intelligence and machine-learning algorithms to perform a deep analysis of aggregated visitor data. This data, when combined with your Adobe Commerce catalog, results in highly engaging, relevant, and personalized experiences for the shopper.

You can create, manage, and deploy recommendations across your store views directly from the Adobe Commerce Admin. If your storefront is implemented using PWA Studio, refer to the PWA documentation. If you use a custom frontend technology such as React or Vue JS, learn how to integrate Product Recommendations into your headless storefront.

For information about installing product recommendations modules, refer to the developer documentation.

Dashboard

On the Admin sidebar, go to Marketing > Promotions > Product Recommendations. The Product Recommendations dashboard appears. From here, you can create a new recommendation or edit an existing recommendation.

This dashboard displays a table of previously configured recommendations (if any) along with the metrics being tracked on those recommendations. You can use the date filter to refine the recommendations by Last 24 hours, Last 7 days, and Last 30 days.

Available recommendation types

See Recommendation Types to learn about the available recommendation types in Adobe Commerce.

Training time for machine learning models

Some recommendation types use behavioral data from your shoppers to train machine learning models that build personalized recommendations. Other recommendation types use catalog data only and do not use any behavioral data. If you want to start quickly, you can use the following, catalog-only recommendation types:

• More like this
• Visual similarity

So when can you start using recommendation types that use behavioral data? It depends. This is referred to as the Cold Start problem.

The Cold Start problem is a measure of how much time a model needs to train before it can be considered high quality. In product recommendations, it translates to waiting for Adobe Sensei to train its machine learning models before deploying recommendation units on your site. The more data these models have, the more accurate and useful the recommendations will be. Collecting this data takes time and will vary based on traffic volume. Because this data can be collected only on a production site, it is in your best interest to deploy data collection there as early as possible. You can do this by installing and configuring the magento/production-recommendations module.

The following table provides some general guidance for the amount of time it takes to collect enough data for each recommendation type:

Other variables that can impact the time needed to train:

• Higher traffic volume contributes to faster learning
• Some recommendation types train faster than others
• Adobe Commerce recomputes behavioral data every four hours. While you can technically deploy your recommendation units at that time, know that the recommendations will become more accurate the longer they are used on your site.

While data is collected on production and machine learning models are trained, you can implement the remaining tasks necessary to deploy recommendations to your storefront. By the time you have finished testing and configuring recommendations, the machine learning models will have collected and computed enough data to build relevant recommendations thus allowing you to deploy the recommendations to your storefront.

Until there is enough training data collected, Adobe Commerce uses backup recommendations to populate your recommendation units.

Backup recommendations

If there is not sufficient input data to provide all requested recommendation items in a unit, Adobe Commerce provides backup recommendations to fill those items.

For example, if you deploy the Recommended for you recommendation type to your homepage, a first time shopper on your site will not have generated enough behavioral data to accurately recommended personalized products. In this case, Adobe Commerce will surface items based on the Most viewed recommendation type to this shopper.

The following recommendation types will fallback to Most viewed if there is not sufficient input data collected:

• Recommended for you

• Viewed this, viewed that

• Viewed this, bought that

• Bought this, bought that

• Trending

• Conversion (view to purchase)

• Conversion (view to cart)

Product recommendations placement

You can place the recommendations in one of the following page locations.

Some storefront pages restrict where you can place the recommendations. Refer to the table below for more information.

• At the top of main content - Recommendations appear above the main content area just below the top navigation bar.
• At the bottom of main content (default) - Recommendations appear below the main content area and before any other content blocks on the page, like Related Products.

Supported recommendations for each storefront page

The following table lists the storefront pages, where you can place the recommendations, and the recommendation types allowed on that page.

B2B support in Product Recommendations

B2B storefronts often require complex logic that dictates product visibility and pricing for each shopper or customer group. Product Recommendations now support this functionality by honoring category permissions, shared catalogs, and customer group-specific pricing. For example, if you have hidden certain categories from your retail customer segment, then a shopper in that segment would not be shown recommendations for products in those categories. Also, when you define a shared catalog for specific customer groups and companies, those shoppers will see recommendations only for products they can access. All recommended products will reflect correct customer group-specific price based on each shopper’s customer group.