Nosto's score insights allow you to understand the different contributing factors on your product listings and directly optimise to your business needs. Score insights are explaining the influence of merchandising rules and other factors on search, category and recommendation product listings. By understanding the impact of different settings, you can instantly take related actions like editing searchable fields or optimising your existing merchandising rules.

In this article, we have a close look on how to get started with the tool, walking you through the different capabilities that Nosto's score insights is offering to you.

This list shows you all merchandising rules that have an impact on the currently shown product listing. Checking this list allows you to get a high-level understanding of influencing factors.

There is a noticeable difference between analysing category listings and search results to Recommendations.

Please keep this important fact in mind as you read through the following guide: it considers both factors.

This percentage shown in the upper right corner of the product card reflects the number that determines the position of a product in a certain listing. The top ranked product always has a score of 100% and all other products have a score that reflects the difference to the top ranked product. This difference allows you to understand how much change would be needed to move products up or down in the listing.

Search: The overall relevance can be determined by two main influencing factors - on the one hand there are the different merchandising rules that are applied and on the other hand there is the score contribution from keyword matches for search results. The relevance bar, that is shown in each product card, shows the composition of the products score highlighting the impact of merchandising rules (blue) and keyword matches (green) on a specific product.

Recommendations: For Recommendations the same approach applies, but only with the difference of recommendation relevancy replacing Search's keyword matching.

You can get a more granular analysis by looking into the score details.

By clicking on "View score detail", you can access the more granular analysis of the score calculation. These insights might be valuable for you when searching for ways to understand and optimise your search results.

As already mentioned, the score bar shows the allocated impact between merchandising rules and keyword matches, or recommendation's relevancy. The grey space in the bar reflects the difference between the top ranked product and the currently viewed product.

When hovering over the different sections in the bar you can see from which rule the score contribution derives.

This section is split into two main sections for separate analysis:

The overall contribution of all merchandising rules is shown as the green indicator next to the heading.

All contributing rules are listed here - and within every rule you can see a detailed analysis of the score boosts from this rule. By clicking on the rule name you directly jump to the rules editing menu if you need to do any adjustments.

The following information pieces are listed:

Please note that demotion rules lead to a promotion of all products that are not containing the specified field value. E.g. a demotion of discounted products will in fact lead to a promotion of all not discounted products. Therefore you will find the result of demotion rules as promoting contributions at the respective products.

In the Keywords section you can get detailed information and answers to the question why and where a query matched with your product data.

The overall contribution of keyword matches is shown as the green indicator next to the heading.

All matches between the query and your product data are listed - the following information will help you to understand the origin of a match and define change actions if needed:

When viewing the score insights preview for recommendations, it’s important to recognize that it reflects the “most popular products” recommendation type, based on product views. In this preview, popularity—measured by views—determines relevancy. However, for recommendation types like cross-sellers, the relevancy score incorporates a more sophisticated calculation of how closely products are related.

Recommendation algorithms adjust the output, what products are recommended dynamically based on various factors, such as the page load context (e.g., which product is being viewed or added to the cart) and the shopper’s earlier browsing behavior. This logic could also be further refined using customized segmentation targeting options. As a result, the preview output would be dependent on who is browsing, which products they are interacting with, which they interacted before, when they browse, and even where they are located if geographical targeting is applied. Consequently, the preview for recommendations would be influenced by multiple factors which would rarely reflect the experience of the admin UI user who generates the rule.

To simplify the score insights preview for users, "most viewed products" are used, providing a solid foundation for adjusting merchandising rules. While the balance between merchandising and relevancy remains constant, the specific factors that influence the relevancy score itself vary depending on the recommendation type used.

For more information, refer to the Recommendation Glossary, which explains the different recommendation types and how relevancy is defined in each context.

To optimise your product listings, you might want to take some actions. We recommend checking the following suggestions: