Thin content: pages that rank for nothing because they answer nothing

Judge the page by the unfinished task

Read the query, then the page, and ask what the visitor still has to look up elsewhere. If the answer is the actual decision — price range, fit, process, constraint, or next step — the page is thin even at a thousand words.

Word-count targets create the opposite problem: padded pages that repeat a definition so they look substantial to a template, not to a reader.

Improve, merge, or remove

A thin page with a real query and a real service behind it should be expanded with the missing decision detail. A thin page that duplicates a stronger URL should be merged. A page that exists only to catch a keyword variant should be removed or noindexed.

Keep an inventory: URL, target query, unique information, and the action you chose. Cleanup without a record tends to get undone by the next publishing sprint.

• Add the fact only this organisation can state.
• Merge pages that share one answer.
• Remove doorway pages that swap a place name and nothing else.

Do not manufacture depth

Adding a generic FAQ, a city list, or an AI paragraph that could sit on any competitor’s site does not fix thinness. Depth is specific: a constraint, a worked example, a boundary, or a proof point.

After the rewrite, check whether the page now deserves internal links from the guide that introduces the topic. Orphaned improvements rarely get recrawled quickly.

How many words should a page have to avoid being thin?

There is no reliable minimum. A short page that answers a narrow question can be complete. A long page that never gets to fit, scope, or the next step is still thin.

Should thin tag or archive pages be deleted?

If they do not answer a distinct query, stop indexing them. Delete them when nothing internal depends on the URL; otherwise redirect to the closest useful hub.