Skip to main content
“Vector Search” is a search method that doesn’t rely on matching words to a query. After using machine learning on a large dataset, documents are associated with vectors, which are a series of numbers, and then the search query checks those vectors to find the most similar results.This provides the benefit of retrieving results based on the meaning of the search query, rather than simply matching words or phrases, also referred to as a semantic search. “Vector Search” is a search method that doesn’t rely on matching words to a query. After using machine learning on a large dataset, documents are associated with vectors, which are a series of numbers, and then the search query checks those vectors to find the most similar results. This provides the benefit of retrieving results based on the meaning of the search query, rather than simply matching words or phrases, also referred to as a semantic search. When “Vector Search” is enabled, it will allow users to select a “Search Mode” while configuring Site Search. Those “Modes” are:“Standard”, which is basic “Keyword” search.“Vector Only”, where “Keyword” search techniques are not used to return results.And “Hybrid”, where “Keyword” and “Vector Search” techniques are combined to return results. When “Vector Search” is enabled, it will allow users to select a “Search Mode” while configuring Site Search. Those “Modes” are: “Standard”, which is basic “Keyword” search. “Vector Only”, where “Keyword” search techniques are not used to return results. And “Hybrid”, where “Keyword” and “Vector Search” techniques are combined to return results. When using “Hybrid Search”, there are two ways to handle how results are ranked: “Blend” or “Boost”.“Blend Mode” is the default. It levels the playing field, bringing high-scoring semantic results right to the top alongside keyword matches. It’s perfect for broadening discovery.“Boost Mode”, on the other hand, is more conservative. it only amplifies the score of products that appear in both search types. Think of it as a way to “validate” keyword results using vector intelligence, giving an extra push to the items that satisfy both “word-matching” and “meaning.” When using hybrid search, there are two ways to handle how results are ranked Blend or boost. Blend mode is the default. It levels the playing field bringing high-scoring semantic results, right to the top alongside keyword matches. It’s perfect for broadening discovery. Boost mode. On the other hand is more conservative, it only amplifies, the score of products that appear in both search types. Think of it as a way to validate keyword results. Using Vector intelligence, giving an extra push to the items that satisfy both word matching and meaning To enable and manage “Vector Search”, in the left menu, on the “MAIN” tab, click “Search”, then click “Configurations”. To enable and manage “Vector Search”, in the left menu, on the “MAIN” tab, click “Search”, then click “Configurations”. in the “Search Configurations” menu, we can select an existing “Configuration”, or click the “Create New Configuration” button on the top right to create a new one. in the “Search Configurations” menu, we can select an existing “Configuration”, or click the “Create New Configuration” button on the top right to create a new one. On the “Site Search” tab of the “Search Configurations” page, the “Vector Search” section is located on the right-hand side, next to “Settings” and “Spell Correction”. All options are “disabled” until we click the “Enable Vector Search” toggle.We’ll discuss each setting in this section. On the “Site Search” tab of the “Search Configurations” page, the “Vector Search” section is located on the right-hand side, next to “Settings” and “Spell Correction”. All options are “disabled” until we click the “Enable Vector Search” toggle. We’ll discuss each setting in this section. “Search Mode” determines how the system balances “Keywords” and “Semantic” logic. We can choose “Hybrid” or “Vector Only”. For most implementations, a “Hybrid” approach is recommended. “Search Mode” determines how the system balances “Keywords” and “Semantic” logic. We can choose “Hybrid” or “Vector Only”. For most implementations, a “Hybrid” approach is recommended. “Results Returned” sets the maximum number of “Vector-based” search results returned with each query. The range is from “1” to “10”, with the default being “10”. Results returned sets the maximum number of vector-based search results returned with each query. The ranges from 1 to 10 with the default being 10. “Distance Threshold” is a configurable parameter that defines how “conceptually similar” a product must be to the query to be considered a candidate for “multiply behavior”. “0.7” is the default, and results below this threshold have little impact. “Distance Threshold” is a configurable parameter that defines how “conceptually similar” a product must be to the query to be considered a candidate for “multiply behavior”. “0.7” is the default, and results below this threshold have little impact. When “Filter by Threshold” is toggled “On”, it constrains the search results to only products that “meet” or “exceed” the set “Distance Threshold”. When “Filter by Threshold” is toggled “On”, it constrains the search results to only products that “meet” or “exceed” the set “Distance Threshold”. “Hybrid Mode” only appears if “Hybrid” was selected in the “Search Mode” field above. It defines how “Keyword” and “Vector” results are merged.As a reminder, “Blend Mode” is the default mode that brings “Semantic” results meeting the threshold into the same scoring scale as top keyword results.“Boost Mode” amplifies the ranking of products that appear in both the standard “Keyword” search and the “Vector” search. “Hybrid Mode” only appears if “Hybrid” was selected in the “Search Mode” field above. It defines how “Keyword” and “Vector” results are merged. As a reminder, “Blend Mode” is the default mode that brings “Semantic” results meeting the threshold into the same scoring scale as top keyword results. “Boost Mode” amplifies the ranking of products that appear in both the standard “Keyword” search and the “Vector” search. “Multiplier” controls the “weight” or “impact” Vector results have on the final ranking of products. The range is from “0” to “5”, with “1” being the default. Multiplier controls the waiter impact Vector results. Have on the final ranking of products. The ranges from 0 to 5 with 1 being the default. When “Boost Vector Results” is enabled, the system applies existing standard “Site Boost” functions to both “Vector” and “Keyword” results. If disabled, boosts only apply to “Keyword-based” results. When “Boost Vector Results” is enabled, the system applies existing standard “Site Boost” functions to both “Vector” and “Keyword” results. If disabled, boosts only apply to “Keyword-based” results. When we have configured all “Vector Search” settings as needed, we’ll click the “Save” button on the top right. When we have configured all “Vector Search” settings as needed, we’ll click the “Save” button on the top right.

Want to learn more?

Vector Search

Learn more about this topic

Search Types

Learn more about this topic