Noindex vs Nofollow: Which Pages to Hide From Google

Three controls. Three different jobs. Most owners use them like synonyms, and the result is a site where the thank-you page ranks, the internal search results eat crawl budget, and the page you tried hardest to hide is still sitting in Google's index months later.

This is a solvable problem in an afternoon. Here is the decision table, the one mistake that traps almost everyone, and the three-number audit that tells you how much junk Google has indexed.

What is the actual difference between noindex, nofollow and disallow?

One controls indexing, one controls a single link, and one controls crawling — and only the first one keeps a page out of search results. Google's robots meta tag documentation defines noindex as: 'Do not show this page, media, or resource in search results.' Nofollow is defined separately as: 'Do not follow the links on this page.' Robots.txt disallow does neither — it stops the crawl.

The verdict: if the goal is 'this page must not show up in Google,' the answer is always noindex — never robots.txt, never nofollow. Google's own robots.txt introduction says it plainly: robots.txt 'is not a mechanism for keeping a web page out of Google.'

Robots.txt is a crawl-budget and server-load tool. Use it to stop Googlebot from hammering faceted URLs or a session-ID trap. Do not use it as a privacy control or an index control.

Why does noindex plus robots.txt disallow keep a page indexed forever?

Because the noindex tag lives inside the page, and you just told Google it is not allowed to open the page. Google's block indexing documentation states it directly: 'For the noindex rule to be effective, the page or resource must not be blocked by a robots.txt file, and it has to be otherwise accessible to the crawler.'

The same page continues: 'If the page is blocked by a robots.txt file or the crawler can't access the page, the crawler will never see the noindex rule, and the page can still appear in search results.'

This is the belt-and-braces instinct backfiring. The owner adds noindex, then adds a Disallow line 'to be safe.' The Disallow wins the race. The noindex is now a note taped to the inside of a locked door.

Worse, a disallowed URL can get indexed on link signals alone. Google's robots.txt docs: 'we might still find and index a disallowed URL if it is linked from other places on the web. As a result, the URL address and, potentially, other publicly available information such as anchor text in links to the page can still appear in Google Search results.' That is where the ugly 'No information is available for this page' result comes from.

The fix is a sequence, not a switch. Remove the Disallow. Leave the noindex in place. Let Google recrawl the page and read the tag. Only after the URL drops out of the index — check it in the URL Inspection tool, not by eyeballing a site: search — can you re-add a Disallow if you want to save the crawl.

Which pages on a service site should be noindexed?

On a typical 40-page service business site, roughly 6 to 10 page types deserve noindex — and almost all of them are generated automatically by your CMS rather than written by a human. Nobody decided to publish 180 tag archives. WordPress just did it.

Two of those rows are the ones that actually cost money. The thank-you page corrupts your conversion data. The staging subdomain competes with your live site for your own brand terms. Fix those first, then work down the list. If you are auditing this from scratch, our technical SEO audit process for service businesses walks the same crawl in order.

Note the X-Robots-Tag rows. Google's robots meta tag docs describe applying it at the server level — for example, adding 'a noindex, nofollow X-Robots-Tag to the HTTP response for all .PDF files across an entire site.' That is one line in your Apache or NGINX config, and it covers every file you will ever upload.

Should PPC landing pages be indexed at all?

No — noindex your paid landing pages when they duplicate an existing service page, which is nearly always. A PPC page and its matching organic service page target the same query, so indexing both makes Google choose, and Google frequently picks the wrong one: the stripped-down page with no navigation, no internal links, and no supporting content.

Noindex does not stop your ads from running. Google Ads does not need the page in the organic index to serve it. Quality Score is judged on landing page experience and relevance, not on organic ranking.

The exception is a true one-off: a campaign page with content that exists nowhere else on your site and could earn organic traffic on its own. That is a real page. Give it navigation, give it internal links, and index it. If it is 'the plumbing page with the phone number bigger,' noindex it and send the organic intent to the plumbing and HVAC service page you already rank with.

This is also why we run search and paid as one program rather than two silos. Two teams that never talk are how you end up with six indexable variants of the same page bidding against each other in the organic results.

When should you use rel=sponsored or rel=ugc instead of nofollow?

Use rel=sponsored on anything you paid for, rel=ugc on anything a visitor typed, and plain nofollow only when neither fits. Google announced the two new attributes on September 10, 2019 and defined them precisely: sponsored is for 'links on your site that were created as part of advertisements, sponsorships or other compensation agreements,' and ugc is 'recommended for links within user generated content, such as comments and forum posts.'

Three practical points from that announcement. First, all three attributes are now hints, not commands: Google says sponsored, ugc and nofollow 'are treated as hints about which links to consider or exclude within Search.' Second, you can stack them — rel='ugc sponsored' is valid. Third, and this is the one that matters for money: if you paid for a link and did not flag it, you are exposed to a link scheme action. Google is blunt that sponsored links must carry sponsored or nofollow.

So yes — nofollow that paid chamber-of-commerce directory listing, or better, mark it sponsored. It still sends referral traffic and it still exists as a citation. It just stops looking like a bought endorsement. The links worth chasing are the ones you did not pay for, which is the whole premise of link building for service businesses.

One more thing that same announcement settled: nofollow is not an indexing control. As of March 1, 2020, nofollow became a hint for crawling and indexing too, and Google wrote that anyone 'depending on nofollow solely to block a page from being indexed (which was never recommended)' should use a real blocking mechanism instead. Nofollowing every internal link to your thank-you page does not hide it. Only noindex does.

How do you tell if your site has index bloat?

Compare three numbers, and if the largest is more than roughly 1.5x the smallest, you have bloat. A 40-page WordPress site routinely ships several hundred indexable URLs once you count tag archives, author archives, date archives, attachment pages, paginated series, and internal search URLs — none of which anyone wrote on purpose.

Attachment pages are the classic WordPress leak. Upload 200 images and older setups will generate 200 standalone URLs, each containing one image and nothing else. Yoast and most modern SEO plugins redirect attachment URLs to the parent post by default now, but sites that upgraded from an old install often still have the setting off. Check it.

Bloat does not carry a named penalty. What it does is spend your crawl allowance on garbage and dilute the topical focus of the site — Google sees 400 URLs about nothing in particular instead of 40 URLs about emergency plumbing in Phoenix. On a fresh domain with no authority, that is the difference between getting your money pages crawled weekly and getting them crawled quarterly.

Which three numbers should match — and why yours don't

Real page inventory, Search Console's indexed count, and your sitemap URL count should land within about 10% of each other. On a CMS-driven service site they usually are nowhere close — the sitemap says 46, Search Console says 310, and the owner can name 38 pages.

The gap has three usual causes. Your sitemap is auto-generated and includes noindexed URLs, which sends Google a contradictory signal — 'crawl this' and 'do not index this' in the same breath. Your CMS is generating archive URLs you never audited. Or you disallowed a section instead of noindexing it, and Google indexed the URLs anyway from internal links.

Work it in this order. Pull the indexed count from Search Console and export the URL list. Sort by page type, not alphabetically — the pattern will be obvious within thirty seconds. Noindex the junk types at the template level, not URL by URL. Strip noindexed URLs out of the sitemap. Then wait: deindexing is a recrawl-speed problem, and on a low-authority site that can take 4 to 8 weeks.

Do not force it with a robots.txt Disallow because it feels faster. You already know what that does. And do not run the URL removal tool expecting a permanent fix — it hides a URL for about six months, then it comes back if the noindex was never readable.

If your indexed count is three times your real page count, the crawl-directive layer is the cheapest fix on your site — it is template work, not content work, and it usually takes a day. That is exactly the ground our technical SEO service covers, and it is the first thing we look at because there is no point ranking a page Google will not crawl. Get my free audit and we will send you the three numbers for your domain.