Contents4
Sufficient consensus
Query expansion adds related terms to a query to improve recall, using synonyms, thesauri, embeddings or terms from top-ranked documents. It is standard in every large system.
It raises the number of documents considered and it shifts which of them win.
What the result pages leave out
These belong to the subject. People ask about them. They are missing from the agreed coverage.
- Drift. An expanded query can move away from what the person meant, and the person never sees the expansion that caused it.
- Whose vocabulary. Expansion sources encode the language of the documents that already rank, which keeps unusual wording out.
- Ambiguity resolved silently. A word with two meanings gets one of them assigned without asking.
- Consequences for publishers. A page can rank for a phrasing it never used, and lose one it did use, without any change on the page.
- The searcher who meant the rare thing. Expansion optimizes for the common reading. People with the uncommon need are quietly reranked out.
What people actually want to know
- Which queries am I ranking for that I never wrote about?
- How does the engine understand my central term when it has two meanings?
- What happens to the searcher whose situation is the rare one?
- Should I name the alternative wordings explicitly, or does the engine handle it?
Name the thing in the words your readers use, including the ones no tool recommends. Expansion works from vocabulary that already exists.