Query classification groups search queries by intent, topic and business role so each useful cluster can be mapped to an appropriate page or action.
Meaning and scope
Query classification is the disciplined grouping of search phrases by what a searcher appears to need, the topic they name and the kind of page that could satisfy that need. Common labels include informational, navigational, commercial investigation and transactional, but labels are only useful when they lead to an editorial or product decision rather than a spreadsheet taxonomy.
How to inspect it
Read the live result set and the query language before assigning a label. A phrase that looks commercial may return guides, local listings or product categories; a low-volume phrase may reveal a precise problem worth covering. Keep the query, locale, device context and evidence date with the decision because search results evolve.
- Clean duplicates, spelling variants and phrases that are outside the site’s actual audience.
- Group phrases around a shared question or task, not only shared words.
- Review a representative live result set and record the likely page format and existing competing intent.
- Map each viable cluster to one maintained page, a planned page with sufficient scope, or an explicit do-not-create state.
- Revisit classifications when inventory, search results or user language changes.
Practical decisions
Use classification to prevent two opposite mistakes: making many thin pages from loosely related phrases, and forcing unlike questions into one page that answers neither well. For a catalogue, check whether the inventory can serve the intent before creating a landing page. For editorial content, state the reader’s decision or task in the brief before drafting.
Limits and mistakes
Intent is an evidence-based interpretation, not a permanent fact attached to a keyword. Do not call a cluster transactional merely because it contains a commercial modifier, and do not use frequency alone to choose priority. Classification supports a page decision; it does not predict rankings or demand.