Customizing Product Sorting

ElasticSuite offers multiple ways to customize the sorting and ranking of your products:

• Using Search Optimizers: Dynamically boost or bury products based on specific rules.
• Manual sorting in categories: Explicitly define the exact position of products within a specific category.
• Manual sorting for search queries: Explicitly define the exact position of products for specific search terms.

Using the Search Optimizers

You can access the optimizer management grid by navigating to ElasticSuite > Optimizers in the Magento admin menu. From here, you can create, update, or delete your search optimizers.

Creating an Optimizer

The main page displays a grid of all your existing optimizers. To add a new one, click the Add New Optimizer button.

You will then be directed to the optimizer creation form, which contains the following general configuration fields:

Parameter	Default value	Description
Enable optimizer	Yes	Toggles whether the optimizer is active and applied to the storefront.
Store View	(Empty)	Defines the scope of the optimizer. An optimizer can only be assigned to a single store view.
Model	Constant	Determines the scoring mechanism used by the optimizer: • Constant: Applies a fixed score multiplier. • Based on attribute value: Applies a dynamic boost proportional to a specific product attribute’s value. • Based on behavioral data: Applies a dynamic boost proportional to a selected behavioral metric (e.g., views, sales).
Optimizer Name	(Empty)	The internal name of the optimizer for administration purposes.
Active From	(Empty)	The start date for the optimizer. Useful for scheduling temporary merchandising events or sales campaigns.
Active To	(Empty)	The end date for the optimizer.
Request Type	(Empty)	Defines the specific storefront context where this optimizer will be applied: • Catalog Product Search: Standard catalog search results page. • Catalog Product Autocomplete: Product suggestions in the autocomplete dropdown. • Category Product View: Standard catalog category navigation pages. • Quick Order Suggest Search: Product results in the quick order suggestion search. • Related Products: Products displayed in the “Related Products” blocks. • Upsell Products: Products displayed in the “Upsell Products” blocks. • Cross-sell Products: Products displayed in the “Cross-sell Products” blocks. • Visitor Products: Products displayed in visitor-specific recommendation blocks.

Rule Configuration

This section features a flexible rule editor allowing you to define the exact conditions products must meet to be affected by the optimizer.

Examples of rules you can create:

• “Color is Blue”: Apply a boost exclusively to products where the “Color” attribute is “Blue”.
• “Only discounted products”: Apply a boost to products currently featuring a special price.
• “Only in stock products”: Push immediately salable products to the top of your product lists.

Constant Optimizer

When using the Constant model, the percentage value provided acts as a direct multiplier on the base relevance score of the targeted products.

• Boosting (Positive Value): A value of 10% will multiply the score of the affected products by 1.10. As a result, these products will rank higher in the affected listings.
• Burying (Negative Value): A value of -40% will multiply the score of the affected products by 0.60 (calculated as 1 - 0.40). As a result, these products will rank lower in the affected listings.

Optimizer based on attribute value

How this works This model allows you to apply a dynamic relevance boost that scales proportionally based on the numeric value of a specific product attribute (such as weight, margin, or a custom score).

Parameter	Default value	Description
Attribute	(Empty)	Select the specific numeric product attribute the boost will be based on.
Boost impact	(Empty)	Defines the mathematical function used to scale the boost. Options are Low (logarithmic), Medium (square root), or High (linear).
Attribute value pre-multiplier	(Empty)	(Formerly named “Boost value” in older versions). A constant multiplier applied to the attribute’s value before the Boost Impact function calculates the final score multiplier.

Calculation Examples Assuming you set an Attribute value pre-multiplier of 2, here is how the final product score multiplier is calculated based on a product’s attribute value (e.g., weight):

• Low Impact log(attribute value * pre-multiplier):
  • Attribute weight = 1kg ➔ score multiplied by log(1 * 2) = 0.30
  • Attribute weight = 3kg ➔ score multiplied by log(3 * 2) = 0.78
  • Attribute weight = 100kg ➔ score multiplied by log(100 * 2) = 2.30
• Medium Impact √(attribute value * pre-multiplier):
  • Attribute weight = 1kg ➔ score multiplied by √(1 * 2) = 1.41
  • Attribute weight = 3kg ➔ score multiplied by √(3 * 2) = 2.45
  • Attribute weight = 100kg ➔ score multiplied by √(100 * 2) = 14.14
• High Impact attribute value * pre-multiplier:
  • Attribute weight = 1kg ➔ score multiplied by 1 * 2 = 2
  • Attribute weight = 3kg ➔ score multiplied by 3 * 2 = 6
  • Attribute weight = 100kg ➔ score multiplied by 100 * 2 = 200