Query understanding is the set of steps that transform a raw query string into a structured, corrected, expanded representation before retrieval happens.
An embedding is a numeric vector representing the meaning of a piece of text, image or product, positioned so that similar things sit close together in the same space.
Vector search retrieves results by finding the stored embeddings nearest to the query's embedding, ranking by geometric distance rather than word overlap.
Semantic search interprets the meaning of a query rather than matching its characters, so results can be relevant without sharing any words with what was typed.
RAG is a pattern where a language model retrieves relevant documents first and generates an answer grounded in them, rather than answering from its training data alone.
A zero-result search is a query that returns nothing, and it is the single most damaging outcome in website search because the visitor reads it as "this site doesn't have it".
Search abandonment is the share of searches after which the visitor takes no meaningful action — no click, no refinement, no conversion — before leaving.
Search conversion rate is the conversion rate of sessions that included at least one search, usually reported alongside the rate for sessions that did not.
An inverted index is a data structure that maps each term to the list of documents containing it, allowing an engine to resolve a keyword query without scanning the corpus.
BM25 is the standard keyword ranking function that scores a document by how often the query's terms appear in it, discounted by how common those terms are across the corpus and by document length.
Synonyms are configured equivalences that let a search engine treat different words as the same concept — sofa and couch, sneakers and trainers, kurta and tunic.