Noindex, Follow: Does It Still Pass Link Equity?

Noindex,follow tells search engines to keep a page out of the results while still following its links. For a short period that is exactly what happens. The catch is that the combination does not stay stable: Google representatives have said that a page which remains noindex for a long time is eventually treated as if it were noindex,nofollow, so the links on it stop counting as links. If you were counting on a noindexed page to quietly pass authority to the pages it links to, that plan has an expiry date.

This article separates the two directives that get mixed up here, noindex and nofollow, explains the four combinations they produce, walks through why the long-term behaviour makes sense, and gives you a practical way to decide what to do with pages you currently have on noindex,follow. It stays on the directive-combination question; how noindex differs from a robots.txt block, or from a canonical tag, is a separate topic.

Two Directives, Two Different Jobs

The robots meta tag carries instructions for search engines, and the two that matter here control unrelated things. noindex answers the question "may this URL appear in search results?" nofollow answers the question "should the links on this page be used?" A page can be any combination of the two, because the answers are independent.

The word follow is the source of most of the confusion. It is not a directive that does anything on its own. Following links is the default behaviour, so a page with no robots meta tag at all is already index,follow. Writing `follow` explicitly documents your intent for the next person reading the template, but it does not add a capability. Likewise `index` is the default, and `all` means the same as index,follow.

That has a consequence people miss: `noindex, follow` is not a special instruction to pass value through the page. It is noindex plus the absence of nofollow. The page is not being granted a privilege; it is simply not being stripped of one. And because it is a default rather than a promise, a search engine is free to change how it treats that default when circumstances change.

The Four Combinations and What Each Does

Every page falls into one of four states. Seeing them side by side makes the long-term problem easier to spot.

• index, follow (the default): the page can rank, and its links are crawled and can pass signals to their targets. No tag is needed.
• index, nofollow: the page can rank, but the links on it are not used as endorsements. This is rare on purpose, for example a page that is entirely user-submitted links you cannot vouch for.
• noindex, follow: the page is dropped from the results, but its links are still crawled and used, at least for a while. This is the combination people reach for on thin but useful navigation pages.
• noindex, nofollow: the page is dropped from the results and its links are ignored. Nothing on it is supposed to influence anything else.

Why a Long-Term Noindex Becomes Nofollow

Google's own guidance on this combination is sparse, but Google spokespeople have described the behaviour clearly enough. A page that has been noindex for a long time is crawled less and less often, and eventually Google treats it as noindex,nofollow. The reasoning is straightforward once you think about how indexing works.

A link is only useful to a search engine if it sits on a page the engine trusts to be part of its picture of the web. A noindexed page is, by your own instruction, not part of that picture. The engine keeps it around at first because noindex is often a mistake or a temporary state, and dropping its links immediately would punish you for a slip. If the directive stays in place month after month, the most sensible reading is that you really do mean it, and that the page is not worth treating as a source of links. The crawl frequency decays, the page is checked rarely, and its outgoing links stop contributing in any meaningful way.

Notice what this is and is not. It is an observed, described behaviour, not a documented switch with a published timer. Nobody outside Google can tell you whether the cutoff is six weeks or six months, and it may differ by page, by site and by how often the URL is otherwise requested. That uncertainty is the real lesson: noindex,follow is reliable for a short window and unreliable beyond it, and you cannot measure where the line is. A strategy that depends on it is depending on an unspecified behaviour.

What This Means for Link Equity Plans

The pattern that goes wrong is predictable. A site owner has thousands of thin URLs, such as tag pages, filtered category views, or internal search results. They add noindex,follow to keep the clutter out of the index while believing the links on those pages will keep feeding product or article pages with authority. For the first few weeks that may well be true. Then the pages age on noindex, get crawled rarely, and the products that were only reachable through them lose a support they never formally had.

Consider a blog with a tag archive. Each tag page lists twenty posts and is set to noindex,follow. Several older posts have no other internal links pointing at them, so the tag pages are their only route in. Once Google starts treating those tag pages as nofollow, those older posts are effectively orphaned as far as link signals go. They may still be crawled through the sitemap, but they are no longer being reinforced by the site's own structure.

The fix is not a cleverer directive. It is to make sure the pages that matter are linked from pages that are indexable and stable: the category page, the hub article, the main navigation, a related-posts block. Treat anything a noindex page links to as if it had no link from there at all, and ask whether the target still has a real path. If it does, the noindex page is irrelevant to its ranking. If it does not, you have an architecture problem that no robots tag can solve.

When Noindex, Follow Is Still the Right Call

None of this makes the combination useless. It is the right choice when the noindex state is short or when the links on the page are not load-bearing anyway. Good uses include:

• A page in transition. A section you are rebuilding, or a launch page that should stay out of the results for a few weeks. The state is expected to end, so the long-term decay never arrives.
• Utility pages whose links are duplicated elsewhere. Internal search results, sort-order variants and session-based views usually link to items that are already linked from proper category pages. Losing their links costs nothing.
• Deep pagination you chose to keep out of the index. The links matter for discovery of the items, but if those items are also in a sitemap and linked from category hubs, the paginated pages are not the only path.
• Pages that must exist for users but should not rank. Thank-you pages, account areas and print views are noindexed because they are useless as landing pages, not because their outbound links carry value.

Noindex, Nofollow: When to Say It Outright

If the links on a page genuinely should not count, say so. Adding nofollow to a noindex page removes the ambiguity: you are stating what a search engine would eventually decide anyway, and you are not relying on its patience. This is sensible for pages such as staging copies that briefly end up public, login-adjacent pages, and pages that aggregate outbound links you would not vouch for.

The trade-off is that you lose the short-term benefit of the follow default, which matters only if the page is a temporary bridge to content that has no other path. In that case, fix the path rather than preserving the bridge.

Keep in mind that page-level nofollow applies to every link on the page, including internal ones. It is a blunt tool. If only some links should be discounted, use the rel attribute on those individual links instead and leave the page-level directive alone.

How to Check What a Page Is Actually Sending

Directives can arrive from two places: the robots meta tag in the HTML head, and the X-Robots-Tag HTTP response header. The header is the only option for non-HTML files such as PDFs, and it is often set at the server or CDN layer, which is why a page can carry a noindex that nobody can find in the template. When the two conflict, the more restrictive directive wins, so a stray header can quietly override a clean-looking tag.

To audit a page, fetch it the way a crawler would and read both the response headers and the rendered HTML. Then use the URL Inspection tool in Search Console to confirm what Google saw. It reports whether indexing is allowed and which directive blocked it. For a whole site, crawl it with a tool that exports the meta robots and X-Robots-Tag columns, then filter for every URL carrying noindex and sort by how many internal links point to it.

The second half of that audit is the one that catches the real damage. For each noindex URL, list the pages it links to and check whether each target has at least one link from an indexable page. Any target that depends solely on noindex pages is the one that will suffer when the follow default stops being honoured.

One more trap: noindex only works if the page can be crawled. If the same URL is blocked in robots.txt, the search engine never sees the tag, so the directive does nothing. Keep the two mechanisms from contradicting each other.

A Decision Path for Pages You Have Now

When you inherit a site with a pile of noindex,follow pages, go through them in order rather than changing them all at once.

First, ask whether the page should exist. Many noindexed pages are leftovers that serve nobody. Deleting them, with a 404 or 410 if there is no replacement, or redirecting them if there is, is cleaner than carrying a permanent noindex. A directive that stays forever is a sign the page has no purpose.

Second, if the page has a close equivalent, consider whether it is really a duplicate. A canonical tag pointing at the preferred URL consolidates signals instead of asking the engine to ignore the page, and it is the better tool for variants of the same content.

Third, if the page must remain and must stay out of the index, decide whether its links matter. If they do, route that value through indexable pages. If they do not, set noindex,nofollow and stop thinking about it.

Finally, if the page is temporary, set a reminder to remove the directive. A large share of long-term noindex problems are simply forgotten launch-week settings. A quarterly scan for noindex URLs that have been in that state for a long time costs little and catches them.