Concept
Text or product information is converted to vectors; items that are more similar in terms of content are closer together in that representation space.
Vector Search compares numerical representations of search queries and product information to find substantive agreement, even without an exact word match.
Text or product information is converted to vectors; items that are more similar in terms of content are closer together in that representation space.
With long natural search questions, inspiration-oriented discovery and assortments with rich descriptive product data.
SKUs, article numbers and very precise model names usually require explicit lexical or exact matching.
Vector Search is therefore often combined with other search methods in a hybrid ranking.