Example: Fine-tuning search rule ranking factors using store preview data

In store preview, we can see the relevancy scores that the search engine assigns to each catalog entry listed in search results. By understanding these relevancy scores, we can adjust ranking criteria to reorder search results in a search rule that uses the Change Search Result Order action. We can also use store preview to see the effect of your adjustments on the relevancy scores of catalog entries.

In this example, to create a search rule in the Aurora starter store that promotes cocktail dresses manufactured by Versatil in the search results. Without a search rule, the Versatil cocktail dresses are listed lower in the search results than you would like them to appear. This procedure illustrates how to use store preview to assign an effective search rule to the Versatil brand so that Versatil cocktail dresses are listed higher than other brands in the search results.

Optional: We can also use ranking rules to demote certain catalog entries to further fine-tune our search results. However, in this example, the ranking action is used only to promote catalog entries in search results.

Procedure

1. Open the Marketing tool.

2. Use store preview to see the relevancy scores that the search engine assigns to cocktail dresses when no search rules are applied:

   1. Select File > Store Preview.

   2. Click Launch Store Preview.

   3. In the search keyword box, type cocktail dresses and click the search icon. The search results display.

   4. To display the relevancy scores for each catalog entry in the search results, click Show Page Information at the top of the store preview window.

      For each cocktail dress, the relevancy score is displayed above the cocktail dress description, as shown in this example:

      Note the following details in the previous screen capture:

      • There is one Versatil cocktail dress listed 12th in the first page of search results. The search engine assigned a relevancy score of 0.18745854 to this cocktail dress. All other Versatil cocktail dresses are listed on subsequent pages of the search results because their relevancy scores are even lower.

      • The Albini cocktail dress that is listed at the top of the search results has a relevancy score of 0.45662335. This score is the highest relevancy score for cocktail dresses.

      To appear higher in the search results than the Albini cocktail dress, the Versatil cocktail dresses must have a relevancy score that is higher than 0.45662335.

      To achieve this result, assign a ranking factor of 35 to promote Versatil cocktail dresses in our search rule. Based on this store preview data, we can see that these adjustments mean that Versatil cocktail dresses will have the highest relevancy scores for cocktail dresses.

3. Create and activate a search rule that uses the Change Search Result Order action to promote cocktail dresses that are manufactured by Versatil by a factor of 35. The search rule looks like this in the search rule builder:

   

   Where a ranking factor of 35 is assigned to promote Versatil cocktail dresses in our search rule.

   Optional: We can also use ranking rules to demote other cocktail dresses to further fine-tune our search results. However, in this example, the ranking action is used only to promote Versatil cocktail dresses.

4. Use store preview to see how the search rule has elevated Versatil cocktail dresses in the search results:

   1. Select File > Store Preview.

   2. Click Launch Store Preview.

   3. In the search keyword box, type cocktail dresses and click the search icon. The search results display.

   4. To display the relevancy scores for each catalog entry in the search results, click Show Page Information at the top of the store preview window.

      The Versatil cocktail dresses are now listed higher in the search results than other cocktail dresses, and they show an increased relevancy score:

      

Results

By analyzing store preview data, we can create search rules that produce the results we want on the storefront.

Related concepts
Search rules

Related tasks
Testing a search rule in store preview