How meta nofollow works
Nofollow appears as one of the values inside a page’s robots meta tag, and it applies to every link on that page at once. It is often paired with noindex, in which case the page is kept out of results and its links are left alone as well. Used by itself, the page can still be indexed; only the onward links are affected.
This is a different instrument from the rel attribute you put on a single link. The rel version marks one link, and it has companions — sponsored for anything paid for, ugc for content other people submitted — that say why the link is not endorsed. The page-level version says nothing about individual links; it simply covers all of them, including your own navigation and footer if those appear on the page.
Why meta nofollow matters
Its honest use is narrow. A page made almost entirely of links other people added, and which you cannot realistically moderate, is a reasonable candidate: an open submission listing, an old forum, an archive you inherited. Declaring that you do not vouch for anything on it is more truthful than marking hundreds of links one at a time.
The rest of its reputation comes from a technique that no longer works. The idea was that blocking some links would push more ranking value into the remaining ones. Google closed that off long ago — the value assigned to a blocked link is not redistributed to its neighbours, it simply goes nowhere. Anyone still proposing internal link sculpting as a service is working from very old advice.
Where meta nofollow goes wrong
The common misunderstanding is treating it as access control. It does not stop a page being discovered, because search engines reach pages through other sites’ links, sitemaps and their own history of your domain. Nor does it hide anything from a person, a scraper or a tool that ignores directives. If something must not be public, it needs a login, not a tag.
The costly error is applying it too broadly. Because it covers the whole page, a template that adds it will also block the menu, the breadcrumbs and every editorial link in the body copy — cutting crawl paths to sections of the site nobody meant to isolate. It is also the wrong response to a single paid link, where marking that one link is both simpler and more accurate.
How to act on it
Reach for the per-link attributes first. Marking a placement as sponsored, or comment links as ugc, tells search engines what kind of link it is rather than just refusing to endorse it, and it leaves the rest of the page working normally. See rel nofollow for how those values differ.
Keep the page-level directive for pages where every link is genuinely someone else’s, and check that no theme or plugin is applying it where you did not intend. It is set in the same place as the rest of your meta robots instructions, so a crawl that lists directives across the site will find any stray ones quickly.
