Define search preferences

Last updated:
Edit on GitHub

This topic describes how to add product attributes to search and define search preferences.

Prerequisites

  1. If you are new to the Search preferences section, you might want to start with Best practices: Promote products with search preferences.

  2. Create a product attribute for the search preference that you are going to create.

  3. Add the product attribute to a product by creating a product or by editing a product.

  4. To start working with filter preferences, go to Merchandising > Search Preferences.

Review the reference information before you start, or look up the necessary information as you go through the process.

Add product attributes to search and define search preferences

  1. On the Search Preferences page, click Add attribute to search.
  2. On the Add Attribute to Search page, enter an ATTRIBUTE KEY.
  3. Select Yes or No for the following:
    • FULL TEXT
    • FULL TEXT BOOSTED
    • SUGGESTION TERMS
    • COMPLETION TERMS
  4. Click Save. This opens the Search Preferences page with a success message displayed.
  5. Optional: To start using the created search attribute, click Synchronize search preferences. This refreshes the page with a success message displayed.

Reference information: Add product attributes to search and define search preferences

This section describes attributes you see and enter when adding product attributes to search and defining search preferences for them.

ATTRIBUTE KEY

A product attribute to add to search. The values of the product attribute are used to search products by. For example, you add the color attribute to search. One of the values of the attribute is green. When a customer searches for green, the search returns all the products with this attribute value assigned.

Enabling search preferences for attributes with numeric or yes and no values results in a huge list of search results because these values frequently appear in SKUs, product names, and descriptions. Additionally, in search results, SKUs, product names, and descriptions are always displayed higher.

FULL TEXT

Defines if all words in search terms are to be searched. The results that match all the words in a phrase are displayed higher than those matching parts of the phrase.

For example, FULL TEXT is enabled for the following attribute key-value pairs:

Key Value
focus_adjustment Auto/Manual
flash Manual

With FULL TEXT enabled, when a customer searches for Auto/Manual, the products with Auto/Manual and Manual values assigned are both displayed in the results. However, the products with Auto/Manual are displayed on top.

FULL TEXT BOOSTED

Defines if attribute values of an attribute are to receive a higher relevance than the same attribute values of other attributes.

For example, the white-balance and light_exposure_modes attributes have the manual attribute value. The white-balance attribute is assigned to Canon LEGRIA HF R606, and light_exposure_modes attribute is assigned to Sony Cyber-shot DSC-W830. You enable FULL TEXT BOOSTED only for light_exposure_modes. When a customer searches for manual, Sony Cyber-shot DSC-W830 appears higher in the list of results than the product with the white-balance attribute. Full text boosted

If FULL TEXT BOOSTED is enabled for two attributes with the same value, in the search results, the order of the respective products is defined by Elasticsearch. You can improve this by customizing its analyzers.

SUGGESTION TERMS

Defines if alternative search terms are to be provided if a search item is misspelled. For example, the storage_media attribute has the SSD and Flash values. With SUGGESTION TERMS enabled, when a customer searches for flashs, the search results page contains a a suggested search term flash. Include for suggestion

COMPLETION TERMS

Defines if auto-completion suggestions should appear when customers search for the values of this attribute.

For example, the storage_media attribute has SSD and Flash values. With COMPLETION TERMS enabled, when a customer enters fla in the search field, the search term is autocompleted with sh, and there is a list of suggested terms in the search results. Completion terms

Also, it does not make much sense to activate search preferences for attributes with the numeric and Yes/No values. As numbers may occur not only in attributes but in product SKUs, names and descriptions (which are actually ranked higher than attributes in search results), therefore the probability that a user will find what they were looking for is low, but the list of search results will be huge, and the search term will be present in multiple places. Besides, it is very unlikely that users will be searching for an attribute with a numeric value or the Yes/No values.