How to Noindex Pages in WordPress

Noindex is the instruction that keeps a page out of Google's index while leaving it alive for users — the right tool for thank-you pages, thin archives, internal search results and the whole category of exists-but-shouldn't-compete content — and WordPress makes applying it a checkbox... in several different places, which is where the confusion lives. Here's every method, when each applies, and the noindex mistakes that make it WordPress's second-most-dangerous checkbox after the visibility setting.

The methods, by scope

1. Per-page: the SEO plugin's advanced tab — the standard path: edit the post/page, Advanced (Yoast) or the robots settings (Rank Math), "No" to indexing — the plugin renders the meta robots noindex tag. This is the tool for the one-off exclusions: thank-you and confirmation pages, gated-content stubs, campaign landings that duplicate organic pages, test pages that must live.
2. Per-type and per-taxonomy: the plugin's Search Appearance settings — the indexation map: whole content types (attachment pages), taxonomies (tags, formats), and archive classes (date, author) excluded wholesale — set once, governing everything current and future in the class.
3. The whole site: Settings → Reading's discourage checkbox — legitimate on staging (behind auth, as the second layer) and catastrophic in production; the standing deployment assertion exists because of it.
4. The header alternative: X-Robots-Tag — server-level noindex for non-HTML files (PDFs, feeds) the meta tag can't carry; host config or plugin-emitted, the specialist's tool.

The rules that prevent the classic mistakes

Noindex requires crawlability — the standing mechanics: a page blocked in robots.txt can't have its noindex read; to de-index something, let it be crawled wearing the tag (the block-first instinct freezes pages in the index). Noindex is not the duplicate tool — true variants want canonicals (consolidation), not exclusion (which discards their signals), per the decision grid; noindex is for pages that are thin or private-ish, not copies. Long-noindexed pages drift to less-crawled — expected and fine for the excluded class; the gotcha is links: a heavily-noindexed nav section still passes link value, but don't build equity plans through pages you've excluded. And audit the map annually: the indexation review catches both directions — money pages accidentally noindexed (the plugin checkbox ticked during drafting and forgotten — a real and recurring finding) and junk classes indexing that shouldn't (new plugin post types arriving public by default).

Diagnosing noindex problems

The symptoms and their reads in Search Console: "Excluded by noindex" on pages that should rank — the accidental-checkbox hunt (URL Inspection shows the tag; the plugin's per-page setting or a type-level rule is the source — check both scopes); pages you noindexed still showing in results — recrawl lag (days to weeks by page importance; request indexing to accelerate the tag's reading) or the blocked-so-unread mechanics above; and the site-wide horror (everything excluded) — the Reading checkbox or a plugin conflict emitting global noindex, found in one view-source. Every diagnosis ends the same way: one emitter, deliberate scope, verified rendering — the silo's whole governance doctrine in miniature.

Frequently asked questions

Should I noindex low-traffic old posts?

Not as a reflex — low traffic isn't the criterion; value is: the audit's verdicts (update, merge, prune) beat blanket exclusion, and noindexing content that could be refreshed throws away its accumulated signals. Noindex is for pages that should never compete, not pages currently losing.

Does noindexing thin pages improve the rest of the site?

At scale, observably — removing a thin-page flood from the index concentrates the quality aggregate on what remains, per the pruning evidence. One thin page noindexed does nothing measurable; the archive-and-attachment classes done wholesale is where the effect lives.

Noindex or delete for content we're retiring?

Delete-with-redirect (relevant successor) or honest 410 for genuinely dead content — noindex keeps the page consuming maintenance and crawl attention for no one; it's the tool for pages that must stay usable, not pages that should end. The retirements feed the same goal as the exclusions: an index footprint that's all signal — beneath which the ranking contest stays the usual pair, content and authority (our signal).