What is noindex? When to use it and when not to
·4 min read
Noindex is a rule that tells search engines not to show a page in their results. Google still crawls the page, reads the rule, and drops the URL from its index, which also keeps it out of AI Overviews and AI Mode. Use it on pages people should reach by clicking through your site but never land on from a search, and nowhere else.
Noindex is easy to add and quiet once it's there, so it spreads to pages that should rank. Most of this post is about telling the two apart.
Noindex meaning, in one tag
The noindex tag is a robots meta tag in the page's <head>, or the same rule sent as an X-Robots-Tag response header for files like PDFs.
<meta name="robots" content="noindex">Two conditions make it work. The page must be crawlable, because a robots.txt block stops Google from ever reading the tag. And the rule must be in the HTML your server sends, not added later by script. The full syntax, server configs and deindexing steps are in noindex in robots.txt.
Which pages should be noindexed
A page deserves noindex when a searcher who landed on it would be confused, or would land somewhere useless. These are the usual candidates.
Thank-you and confirmation pages. They only make sense after a form submit or a purchase.
Internal search results. Every query on your site search makes a new URL with thin, shifting content. Noindex them.
Thin tag and archive pages. A tag with two posts on it adds nothing a searcher can't find on the posts themselves. Noindex those, but check first that your theme or shop doesn't use tags as real landing pages. Stores often do.
Filter and sort variants. Google's pagination and incremental loading guide suggests noindex for filtered or sorted versions of a list you don't want in search.
Staging and test copies. Noindex works, but a password works better. A forgotten password at launch gets noticed in minutes. A forgotten noindex can sit for weeks.
Gated or duplicate versions you can't canonicalize. A print view or a members-only copy of a public page is a candidate when a canonical tag isn't possible.
Pages that should not get noindex
Pages you want found should never carry noindex, and a few kinds get caught by accident.
Paginated category pages are the common one. Page 2 of a category is how Google reaches products 25 through 48. Google's guide says to give each page in a series its own canonical URL rather than pointing them all at page 1, and it does not suggest noindexing the series. Leave them indexable.
Duplicates within your own site are the other. Google's duplicate URL guide says it doesn't recommend noindex to stop a page being chosen as canonical, because noindex blocks the page from Search completely. If two URLs show the same content, a canonical tag passes the signals to the one you want. Noindex just throws one away.
And noindex is the wrong tool for pages you want indexed but not quoted. That job belongs to nosnippet and max-snippet.
Noindex vs canonical vs 404 vs password
Pick by what should happen to the URL and its visitors.
| Situation | Use | What happens |
|---|---|---|
| Page is useful to visitors, useless from search | noindex |
Stays live, leaves the index |
| Two URLs show the same content | Canonical tag | One URL ranks, the other's signals move to it |
| Page is gone for good | 404 or 410 | Google drops it after recrawling |
| Page moved | 301 redirect | Visitors and signals follow |
| Nobody outside should see it | Password or login | Crawlers and people both stopped |
The test I use is whether a human would miss the page. If nobody would, delete it and return a 404 or 410. If people use it but shouldn't arrive cold from Google, noindex it.
Noindex, follow: does it still pass links?
Not for long. noindex, follow asks Google to drop the page but keep following its links, and follow is the default anyway, so writing it changes nothing.
The catch is time. In a December 2017 Webmaster Central hangout, Google's John Mueller said a page kept on noindex long term eventually gets treated like noindex, nofollow, because Google drops it entirely and stops following its links, as Search Engine Roundtable reported. He gave no timeframe.
So don't rely on a noindexed hub to pass link equity to anything. If a product or article is reachable only through noindexed tag pages, link to it from an indexable category, a sitemap entry and related posts too.
Does noindex affect AI search?
Yes, for Google. Its AI features guide says a page must be indexed and eligible to show with a snippet to appear as a supporting link, and lists noindex alongside nosnippet, data-nosnippet and max-snippet as the ways to limit what Search shows. A noindexed page cannot be a source in AI Overviews or AI Mode.
One stray rule on a pillar page now costs you the ranking and the AI citation together. If you find a page excluded for this reason in Search Console, excluded by noindex tag walks through finding which system set it.
Check a page for noindex
Our Indexing & Canonical Checker fetches a URL, follows redirects, and reads the HTTP status, the meta robots tag and the X-Robots-Tag header separately, so you can see which layer set a noindex. It reads robots.txt for Googlebot and warns when a Disallow hides the noindex from Google. It also checks the canonical tag, loads its target to confirm that it's indexable, and reads snippet controls for Google and Bing.
A run costs 8 credits. It reads the HTML your server sends without running JavaScript, so a noindex added by script won't show up, and it can't confirm whether Google has the URL in its index. URL Inspection in Search Console does that.