Semantic search
What is semantic search?
Semantic search is search that tries to understand the meaning and intent of a query, and the meaning of content, rather than only matching the exact words. It uses context, synonyms, entities and the relationships between concepts, so a search for “how to fix a dripping tap” can return a page titled “Repairing a leaking faucet”.
Google has moved steadily in this direction: the Knowledge Graph in 2012, the Hummingbird algorithm in 2013, RankBrain in 2015, and language models such as BERT in 2019 and MUM in 2021. AI search systems go further, often using vector embeddings, numerical representations of meaning, to find passages that answer a question even when they share few words with it.
Why semantic search matters for SEO
Semantic search is why exact-match keyword tricks matter less than they used to, and why keyword stuffing doesn’t work. A page ranks because it covers the topic well and satisfies the search intent, not because it repeats a phrase.
It also means one good page can rank for hundreds of related queries, and that covering a topic thoroughly, with the related concepts people expect, helps a page be found for queries you never targeted directly.
How to write for semantic search
- Start from the intent: what does the searcher need to know or do?
- Cover the topic fully, including the related questions and subtopics people expect to see.
- Use natural language and synonyms; don’t force exact-match phrases.
- Group closely related keywords on one page rather than one page per variant (see keyword clustering).
Semantic search FAQ
What is semantic SEO?
Semantic SEO is writing and structuring content around topics, entities and intent rather than single keywords, so search engines understand what the page is about and which queries it answers.
Are LSI keywords part of semantic search?
Not really. Google has said it doesn’t use latent semantic indexing. Covering related concepts naturally helps, but “LSI keywords” as a technique is a myth.
Related SEO terms
An entity is a distinct, identifiable thing, such as a person, place, organisation or product, that search engines and AI models recognise.
The Knowledge Graph is Google’s database of entities and the facts and relationships between them, used to power knowledge panels and answers.
Search intent is the reason behind a search: what the person wants to know, find or do. Matching it is one of the most important parts of SEO.
LSI keywords is an SEO name for related terms. Google says it doesn’t use latent semantic indexing; what matters is covering the topic’s related concepts.
Keyword clustering is grouping keywords that share intent and search results, so each group can be targeted by one page.
A large language model (LLM) is an AI model trained on huge amounts of text to predict and generate language, such as GPT, Gemini or Claude.