Noindex and Nofollow Together: When It Is Right and When It Hurts

This check flags any page whose robots meta tag (or X-Robots-Tag header) carries both noindex and nofollow. That combo pulls the page out of search results and tells crawlers to ignore every link on it — which is exactly right for a handful of pages and quietly damaging on the rest, so the flag is a prompt to confirm you meant it.

What each directive actually does

People conflate these two constantly, so here is the clean split. noindex controls the page: keep this URL out of the index. nofollow at the page level controls the links: don't use any outbound link on this page for discovery or equity. They're independent switches. You can set either, both, or neither.

The row that catches people is the last one. Once the page is a dead end, any URL reachable only through that page loses its discovery path. On a small site that rarely matters; on a large one with orphan-prone sections, it strands pages.

Here is a typical flagged page:

<meta name="robots" content="noindex, nofollow">

Or the header equivalent, which the check catches just the same:

HTTP/1.1 200 OK
X-Robots-Tag: noindex, nofollow

When noindex, nofollow is the right call

There is a real set of pages where you genuinely want the URL hidden and don't care about crawling onward from it: internal search results pages, "thank you" / order-confirmation pages, login and account screens, cart and checkout steps, unsubscribe confirmations, and staging or preview URLs that leaked. None of those should rank, and none of them are navigation hubs you need crawlers to spider through. For those, noindex, nofollow is clean and correct.

When it is a mistake

The damage shows up when nofollow rides along on a page that does host useful internal links. Classic cases: a noindexed category or tag archive that is still the only path to a batch of posts; a noindexed paginated series where ?page=2 onward is reachable only via the paginated links you just told Google to ignore; a "printer-friendly" or filtered view that duplicates a page but still links deep into the site. In all of these, noindex, follow is what you actually wanted — hide the page, keep the crawl paths. Defaulting to nofollow "to be safe" is where sites accidentally orphan content.

How to detect it

1. Screaming Frog: after a crawl, go to the Directives tab and use the filters for Noindex and Nofollow. The URLs appearing in both filters are your noindex, nofollow set. Export it and eyeball whether each one is a legitimately dead-end page or a link hub.
2. GSC URL Inspection: inspect a suspect URL; the "Indexing allowed?" line will read "No: 'noindex' detected." It won't spell out the nofollow, so pair this with a source check to confirm both directives are present.
3. View Source: open the raw HTML (Ctrl+U) and search for name="robots". Watch for two robots tags — a theme default plus a plugin override — because a later tag can flip the directive and the check reads the effective one.
4. Response headers: run a header check (browser Network tab or curl -I) for X-Robots-Tag. A header-level noindex, nofollow is invisible in View Source and traps a lot of people debugging "but the meta tag says index."

How to fix it

Ask one question per flagged page: does anything of value link out from here that Google should keep crawling? If yes, drop the nofollow and keep the page hidden:

<meta name="robots" content="noindex, follow">

If the page is a true dead end (checkout, thank-you, internal search), leave noindex, nofollow as-is — the flag is informational, not an order to change it. When you do edit the directive, make sure the page isn't also blocked in robots.txt; a disallowed URL never gets recrawled, so Google keeps acting on the old directive indefinitely. And if you're setting this at scale via headers, double-check the header rule isn't catching more URL patterns than you intended.

Common mistakes

The most frequent one is a global SEO-plugin setting that stamps noindex, nofollow on an entire post type or archive, when the intent was only to hide the page from results. Another is copy-pasting a "block this page" snippet from a forum that includes nofollow without the author realizing it kills internal link flow. A third: adding nofollow to a noindexed page to "save crawl budget" — on nearly every site that trade is a loss, because you're spending link-graph health to save a trivial amount of crawling.

FAQ

Does nofollow in the robots meta tag also affect a single link?

No — the page-level robots nofollow applies to every link on the page at once. To nofollow one specific link, use rel="nofollow" (or sponsored/ugc) on that anchor instead. The two mechanisms are unrelated despite the shared word.

If a page is noindex, why does the follow part still matter?

Because a noindexed page can still be crawled and still passes discovery and equity through its links when set to follow. The page stays out of results while acting as a working junction in your internal link graph. Kill the follow and it becomes a dead end.

Is noindex, nofollow stronger or "safer" than noindex, follow?

Not safer — just more restrictive. It gives up crawl paths you usually want to keep. Reach for it only on genuine dead-end pages. For everything else, noindex, follow is the sensible default.

Will noindex eventually turn into nofollow on its own?

Google has said that a long-lived noindex page can, over time, be treated as effectively nofollow because it stops recrawling the stale page as often. That's a reason to fix orphaning at the source rather than relying on follow to prop up a page you've hidden for good — link to important URLs from indexable pages too.

How is this different from the canonical + noindex conflict?

This check is about two robots directives on one page. The related canonicalized-URL-is-noindex conflict is about a robots directive fighting a cross-URL canonical — a different and more damaging contradiction.

For the full behavior of both directives, including header-based delivery, see the meta robots and X-Robots-Tag reference and the noindex glossary entry. If you need to generate a correct tag, the meta robots generator spits out the exact syntax.