What noindex is
Noindex is a direct instruction to search engines: do not show this page in results. It can be delivered two ways, and both count:
In the page's HTML, as a robots meta tag:
<meta name="robots" content="noindex" />
In the server's response headers, invisible in the page code:
X-Robots-Tag: noindex
Google honors both. The header variant is the one that catches people out — you can stare at a page's HTML for an hour and never see it.
When it's the right tool
Some pages genuinely shouldn't be in Google: thank-you and checkout pages, internal search results, thin tag archives, login screens, staging copies. Noindex keeps them out cleanly while still letting visitors and crawlers reach them.
When it's a disaster
A noindex left on a page you want found removes that page from search — completely, regardless of how good the content is or how many links point at it. The classic accident: a site is built with noindex everywhere (sensible for a staging environment), launches, and the tag ships to production. Traffic never arrives and nothing looks broken.
If a page that should rank gets no impressions at all, checking for noindex — in both the HTML and the headers — is the first diagnostic.
The robots.txt interaction everyone gets wrong
For Google to see a noindex tag, it must be allowed to crawl the page. If robots.txt blocks the page, Google never reads the HTML, never sees the noindex — and the URL can still appear in results anyway. To remove a page from search: noindex it and leave it crawlable. Google spells this out in its noindex documentation.
Removing one
Delete the meta tag or the header, then let Google recrawl the page. Recrawling can take days to weeks; requesting indexing in Search Console for the affected URL speeds it up.