What is the Search Intent?

Search intent is the – assumed – goal behind a search query: what the user actually seems to be wanting to accomplish, independent of the specific words they used. Classifying intent — informational, navigational, commercial, transactional — allows content to be structured to match what users need rather than just what they typed. It is a significant improvement on keyword-only thinking. It also has a hard boundary.

What intent classification achieves

Intent classification allows you to match content structure, depth, and format to what users are actually trying to do. A query with informational intent needs a different response than one with transactional intent. Understanding this distinction improves relevance, reduces bounce, and aligns content with what search engines reward.

Intent classification works by examining existing search behaviour — what users search, what they click, what they do next. It is, in this sense, a map of expressed needs: needs that have already been articulated in search queries.

The boundary: unexpressable intent

There is a category of need that intent classification cannot reach: the need that cannot yet be expressed as a search query, because the vocabulary to express it precisely doesn’t exist yet.

A practitioner who has identified a pattern in their domain — a recurring problem, a structural gap in existing approaches, a phenomenon they have observed but cannot yet name — cannot search for it. They can search around it using approximate terms. But the precise need remains unaddressed, because the system has no query to match it to.

This is the Semantic Vacuum: intent without addressable vocabulary. No amount of intent classification can surface it, because it has never generated search behaviour that classification tools can observe.

Creating addressable intent

The act of naming a concept — establishing a DefinedTerm with schema, a definition page, and external corroboration — does something unusual: it creates an addressable query where none existed. Once the term exists and is indexed, the intent that previously had no expression gains one. This is not optimisation within existing intent. It is the creation of new intent space.

Frequently asked questions

Is there a fifth intent category for pre-consensus needs?

The standard four-category model doesn’t accommodate pre-consensus intent well. A useful framing is “definitional intent” — the need to establish what something is called before any other search behaviour is possible. This is the first-moment intent that Ignorance Graph positioning aims to own.