Look! Those chocolate truffles have a good price! It’s Valentine’s Day, and everybody is already buying them!

Recommendations shows the top searched items. That’s to say, the most clicked products on the search results page, based on shopper interaction within a defined period.


Recommendations are relevant product suggestions; they aren’t related to the query, but to products that matched previous searches done by other shoppers, instead. They are useful in numerous situations, especially when there are no search results to show. Whether a shopper lands on your shop and hasn’t written any query yet or if a search doesn’t bring back any product results, Recommendations amaze shoppers with product discovery inspirations.

Spot the difference

Notice that Recommendations guide shoppers to specific products. When clicking on a recommendation, the related product description page displays without any further interaction. No need to search or navigate the results. Try to not confuse it with the Popular Searches feature, which presents the top searched queries to trigger a new search.

For example, “off-the-shoulder ditsy floral dress” can refer to a top-clicked specific product that redirects to the product description. But “dress”, “floral dress”, or “off-the-shoulder dress” express a currently top searched query, prompting a new search and displaying relevant results such as “off-the-shoulder ditsy floral dress”, among others.

You can identify Recommendations mostly under labels like Top searched items today.


Remember that other types of suggestions can display hints based on queries, search history, or even product references.

Try Recommendations to...

  • Improve product findability. Guide shoppers to products in your catalogue they don’t know.
  • Enhance shopper experience. Surprise shoppers with new ideas or alternative products.
  • Handle shopper frustration. Suggest products when there’s no query to suggest or no results to show.
  • Speed up the search and discovery process, especially on mobile devices.

The inner workings of Recommendations

Recommendations are based on shoppers’ actions in the search UI. As shoppers perform searches in the commerce store, the Tagging microservices track the clicked products, and the Statistics microservice analyzes the data to identify the most clicked products. For example, if over the past weeks shoppers have clicked on specific beach towels, swimsuits, and flip-flop articles, the same products are identified as valid recommendations.

After a number of days of data collection, the Top Clicked batch process generates the top-clicked product feed and calls the Index microservice to index the processed products in the search engine to make them available as recommendations.

At query time, Empathy Platform Interface X uses the Search microservice to retrieve the list of recommendations to display before, during, or after the search, depending on your implementation.


For a correct performance, make sure that your current search service supports this type of feature.


Explore the interactive map to see how Recommendations relate to the other Empathy Platform features and microservices.

We are cookie-less

This is a cookie-free area. Who needs cookies when you can have trust? Feel free to browse our site and the only cookies you need to worry about are the edible kind.