rel nofollow, sponsored and ugc: which link attribute to use
Which rel value goes on paid, affiliate, comment and untrusted links? Google's rules for nofollow, sponsored and ugc, plus internal links and AI crawlers.
By Crawlito · 9 min read
In short
- Google documents three rel values for outbound links: sponsored for ads and paid placements, ugc for comments and forum posts, and nofollow when neither fits.
- Google says links marked with these values will generally not be followed, and its 2019 announcement describes them as hints rather than commands.
- Google's 2019 announcement says there is no need to change nofollow links you already have; switching paid links to sponsored is optional.
- For links to your own pages, Google says to use a robots.txt disallow rule if you need to prevent a page from being fetched.
- Amazon and Common Crawl say their crawlers honor nofollow, while OpenAI's, Anthropic's and Perplexity's crawler documentation does not mention it.
On this page
- What do nofollow, sponsored and ugc mean in Google's documentation?
- Which value should you use?
- Can you combine values, and does Google obey them?
- Do you need to change old nofollow links?
- Why is nofollow rarely the right tool for internal links?
- What do link attributes do for AI crawlers?
- How to find and audit rel attributes on a page
- Frequently asked questions
- Check your own links
Use rel="sponsored" on paid links and ads, rel="ugc" on links inside comments and forum posts, and rel="nofollow" only when neither fits and you would rather not vouch for the page you link to. You can combine values, and you do not need any of them on ordinary editorial links or on links to your own pages.
These attributes tell Google about your relationship to a link. They do not hide a page, and the AI crawler documentation cited below says little or nothing about them. This guide covers what Google says about each value, how to pick one, whether old links need changing, why internal nofollow is rarely the right tool, and how to audit a page.
What do nofollow, sponsored and ugc mean in Google's documentation?
Google's page on qualifying outbound links says ordinary links need no rel attribute. For other links, you add a value to the <a> element.
| Value | Google says to use it for |
|---|---|
sponsored | Advertisements and paid placements, commonly called paid links |
ugc | User-generated content links, such as comments and forum posts |
nofollow | Cases where the other values do not apply, and you would rather Google not associate your site with the linked page or crawl it from your site |
Google says links marked with these attributes "will generally not be followed". It adds that a linked page can still be crawled if Google finds it another way, such as through a sitemap or links from other sites.
Two details about where these values come from:
nofollowis part of the HTML Standard. The standard defines it as a statement that the original author or publisher of the current document does not endorse the linked document. At the time of writing, the standard's link types section does not listsponsoredorugc. Those two are values Google documents for its own use.relis a set of space-separated keywords. Browsers read it as a list, so one link can carry several values.
Which value should you use?
Start from why the link exists, not from what the target looks like.
| Situation | Value to use | Notes |
|---|---|---|
| An ad, or a placement someone paid you for | sponsored | Google's documented value for paid links |
| An affiliate link where you earn money | sponsored | See the note below; Google's page does not name affiliate links |
| A link in a visitor comment | ugc | Applies to the visitor's link, not your own links in the same article |
| A link in a forum post | ugc | Same rule |
| A link you do not trust, and none of the above applies | nofollow | Google says to use it when other values do not apply |
| A link you chose to include in your own content | None | No attribute is needed |
| A link to another page on your own site | None | Covered in the internal links section below |
What about affiliate links?
The Google pages cited in this post do not use the word affiliate. Google's 2019 announcement says sponsored is for links created as part of advertisements, sponsorships or other compensation agreements. A link you place because you earn a commission is a compensation arrangement, so sponsored is the closest documented fit. Google also accepts nofollow on paid links, and says sponsored is preferred.
How does Google's spam policy connect to paid links?
Google's spam policies list exchanging money for links as link spam when it is done to manipulate search rankings. They also say buying and selling links for advertising and sponsorship is a normal part of the web, and that qualifying those links with nofollow or sponsored keeps them within policy. The same policies list excessive link exchanges and partner pages built only for cross-linking.
The spam policy page names nofollow and sponsored as the paid-link qualifiers and does not mention ugc. For comments and forum posts, Google's link guidance suggests ugc or nofollow, and the qualifying page recommends ugc.
Should every comment link get ugc?
Google's qualifying page recommends marking UGC links with ugc. It also says you might drop the attribute for members who have contributed high-quality content consistently over time. Treat that as a judgment call about specific people, not a default. Google's spam policies list comment spam on blogs and spammy forum posts as examples of user-generated spam, and forum comments with optimized links in the post or signature as an example of link spam.
Can you combine values, and does Google obey them?
You can. Google says to list multiple values separated by spaces or commas, and the HTML standard defines the space-separated form.
<a href="https://partner.example/offer" rel="sponsored">Partner offer</a>
<a href="https://example.org/" rel="ugc">A commenter's link</a>
<a href="https://example.org/" rel="ugc nofollow">A commenter's link, with fallback</a>Google's 2019 announcement says rel="ugc sponsored" is valid, and that nofollow can sit beside the newer values if you want older tools that only know nofollow to still understand the link.
On whether Google obeys them: the announcement describes nofollow, sponsored and ugc as hints about which links to consider or exclude in Search, used with other signals. It says nofollow became a hint for crawling and indexing on March 1, 2020. Google's current page says marked links will "generally" not be followed. The wording leaves room for exceptions, so do not treat a rel value as a guarantee that a link is ignored.
Do you need to change old nofollow links?
No. The 2019 announcement says there is no need to change nofollow links you already have. It suggests moving paid links to sponsored if or when that is convenient, which makes it a tidy-up for next time you edit those pages, not a migration project.
What matters more is the reverse: a paid or compensated link with no attribute at all. That is the case the spam policy is about, so a review of old content is better spent finding unmarked paid links than rewriting working nofollow ones.
Why is nofollow rarely the right tool for internal links?
Google's qualifying page says that for links to your own site, you should use a robots.txt disallow rule if you want to keep a page from being fetched. It says that to keep a page out of the index, you should allow crawling and use noindex. A nofollow on an internal link does neither job reliably, for these reasons:
- A linked page can still be found through your sitemap, other links or external links, so
nofollowdoes not remove it. - It adds nothing for people. Visitors click the link either way.
- It can cut off discovery for crawlers that honor it (see the next section), which leaves the target with fewer routes in. A page that no followed link points to can end up looking like an orphan page to those crawlers.
There is also a page-level version. The nofollow robots rule in a meta tag or X-Robots-Tag header applies to the links on that page as a whole, and none is equivalent to noindex, nofollow. That is a different control from a link attribute, and the practical difference between the two is in robots.txt vs meta robots and noindex, nosnippet and X-Robots-Tag. The general rule for navigation is in broken internal links and crawler discovery: link to every page you want found with a plain <a href>.
What do link attributes do for AI crawlers?
Less is documented here than for Google, so be precise about what is known.
| Operator | What its documentation says about nofollow |
|---|---|
| Amazon | Says Amazonbot, Amzn-SearchBot and Amzn-User respect the link-level rel=nofollow directive |
| Common Crawl | Says CCBot currently honors the nofollow attribute on links embedded in your site, and that the value is not meant for blocking access or preventing indexing |
| OpenAI | The crawler overview page does not mention it |
| Anthropic | The crawler support page does not mention it |
| Perplexity | The bots page does not mention it |
Sources: Amazon's crawler page, Common Crawl's FAQ, OpenAI's crawler page, Anthropic's crawler article and Perplexity's bot guide. Check them again before relying on this table, as operators update their pages.
A missing mention is not a statement that the attribute is ignored. It means the operator has not said either way, so do not plan around it. The pages cited here do not describe any special handling of sponsored or ugc either.
What follows in practice:
- Do not use rel values to control AI use of your content. The documented controls are robots.txt tokens, page-level robots rules and your firewall. See allowing AI search while blocking AI training.
- Avoid
nofollowon your own navigation. The crawlers that say they honor it may skip the link, and others have not said. - Keep paid-link labelling for Google's sake. The documented reason to use
sponsoredis Google's policy, not any AI system.
How to find and audit rel attributes on a page
Check the raw HTML first, because that is what a crawler that does not run JavaScript receives.
curl -s https://example.com/page | grep -oiE '<a [^>]*rel=[^>]*>' | head -50This lists anchor tags that have a rel attribute. It will miss tags that are split across several lines, so treat it as a quick scan, not proof.
Then compare with the rendered page. In your browser's developer console, run:
[...document.querySelectorAll('a[rel]')].map(a => [a.rel, a.href])If a link appears in the console but not in the curl output, a script added it, so check how that script sets rel. Checking JavaScript links for crawler discovery covers that comparison.
Work through this checklist:
- Paid links. Every advertisement, paid placement and affiliate link should carry
sponsored(ornofollow). - Comments and forums. Links in visitor content should carry
ugc, orugc nofollow, unless you have decided to trust specific members. - Internal links. Look for
nofollowinside your navigation, footer, breadcrumbs and related-article blocks. Remove it unless you have a documented reason. - Templates. Check where the attribute comes from. A theme, plugin or component may add it to every link in a region, which can hit editorial links by accident.
- Misspellings.
rel="no-follow"andrel="nofolow"are not documented values.
Frequently asked questions
Is nofollow, sponsored or ugc best for affiliate links?
Use sponsored. Google's 2019 announcement describes it as the value for links created through advertisements, sponsorships or other compensation agreements, and Google says nofollow is also acceptable for paid links.
Does nofollow stop Google from crawling a page?
Not reliably. Google says links marked with these values will generally not be followed, but a page can still be crawled if Google finds it through a sitemap or another site's links. To keep a page out of the index, allow crawling and use noindex.
Should I put nofollow on links to my own pages?
Rarely. For your own pages Google recommends a robots.txt disallow rule when you want to prevent fetching, and noindex when you want a page out of the index. A nofollow on a navigation link can also stop crawlers that honor it from reaching the target.
Do AI search crawlers read sponsored and ugc?
The documentation cited above describes no special handling of either value. Amazon and Common Crawl say their crawlers respect nofollow, and OpenAI, Anthropic and Perplexity do not mention link attributes in the pages cited above.
Can I use more than one rel value?
Yes. Separate values with spaces, as in rel="ugc nofollow". Google also accepts commas.
Check your own links
The free check reads one page the way AI crawlers do, with no account needed, so it is a quick way to see whether your important links are plain HTML in the raw response. If you want the same checks across a whole site, an audit covers up to 500 pages with evidence and prioritized fixes.
Filed under Structure and schema. Topics: internal-links, google, checklist