A blogger opens her inbox and finds an offer she has taken before: a skincare brand will pay to feature its serum, and the draft review is already written. She pastes the brand link into the post, publishes, and moves on. Six months later a competitor files a spam report, and the question that lands back in her inbox is one she never thought to ask. Was that link marked? A single plain <a href> is the difference between a paid mention that follows Google’s rules and one that reads, to Google’s systems, like a bought ranking signal.
That missing attribute is what this covers. Three small values, sponsored, ugc, and nofollow, sit inside the link tag and tell Google how much you stand behind a link you point outward. Choosing the wrong one, or none at all, is one of the quieter ways an otherwise clean site drifts into link-spam territory.
What the three attributes tell Google #
Each value answers a different question about the link, and Google’s own outbound-links documentation is unusually plain about which is which. Mark advertisements or paid placements, what Google calls paid links, with rel="sponsored". Mark user-generated links, such as blog comments and forum posts, with rel="ugc". Use rel="nofollow" when the other two do not apply and, in Google’s words, you would “rather Google not associate your site with, or crawl the linked page from, your site.”
The shared effect is that Google will, in its own phrasing, “generally not follow” a link carrying any of the three. What differs is the reason you are declaring. Sponsored says money changed hands. Ugc says a stranger, not you, put the link there. Nofollow is the catch-all for everything else you do not want to vouch for: a source you cite but distrust, a login page, a competitor you name in passing.
One detail trips people up. You can stack these. Google allows “multiple rel values as a space- or comma-separated list,” so a paid link buried inside a reader comment can legitimately read rel="ugc sponsored". The attribute is a description of the link, not a single-choice dropdown.
From a directive to a hint #
For its first fourteen years, nofollow did one blunt job. When Google, Yahoo, and Microsoft introduced it in January 2005, it was a weapon against comment spam: the attribute told search engines to drop the link and pass no PageRank. It worked. The economics of spamming blog comments collapsed almost overnight, and for well over a decade nofollow stayed a hard instruction.
That changed in September 2019. Google announced sponsored and ugc, and with them a quieter but larger shift: from then on all three attributes would be treated as hints, not directives. They counted as hints for ranking straight away, and from March 1, 2020, nofollow became a hint for crawling and indexing too. A hint means Google reads the attribute, weighs it against everything else it knows about the link, and decides for itself whether to honor it.
This is the part worth sitting with. You are no longer issuing a command. You are filing a declaration Google is free to interpret, and while the attributes still do their job in the ordinary case, the guarantee people remember from 2005 is gone.
Which attribute fits which link #
The decision collapses into a short set of cases. The table below maps the link in front of you to the value Google’s documentation points toward.
| The link is | Attribute | Why |
|---|---|---|
| A paid ad, sponsorship, or affiliate placement | rel="sponsored" |
Money or value was exchanged for the placement |
| Inside a comment, forum post, or profile a user submitted | rel="ugc" |
Someone other than you created it |
| A paid link sitting inside user content | rel="ugc sponsored" |
Both facts are true, so declare both |
| A source you cite but do not endorse, or a page you would rather not pass signal to | rel="nofollow" |
Nothing was paid, no user made it, you still do not vouch |
| A genuine editorial reference you stand behind | no attribute | A followed link is your vote of confidence |
The last row is the one people forget. Qualifying every outbound link “to be safe” strips a site of the ordinary editorial links that make the web legible. A followed link is not a liability. It is the default, and a normal mix of followed and qualified links is what a real site looks like.
The nofollow trap: it does not block crawling #
Here is the assumption that causes the most damage: that a nofollowed link is walled off from Google. It is not. Google’s documentation is direct on the point. A nofollowed page “may be found through other means, such as sitemaps or links from other sites, and thus they may still be crawled.”
Read that against what nofollow became in 2020, a crawling hint rather than a rule, and the conclusion is uncomfortable for anyone using it as a gate. If you nofollow a link to keep a thin or private page out of the index, you have done less than you think. The page can still be discovered, crawled, and indexed through any other path that reaches it.
Keeping a page out of search is a different tool’s job. A noindex robots meta tag on the target page, or an authenticated barrier in front of it, controls indexing. The link attribute controls endorsement and, loosely, crawling from your own site. Confusing the two is how a supposedly blocked page ends up ranking for a query its owner never wanted.
Where this meets Google’s spam policy #
The attributes are not decoration. They are how Google separates normal advertising from link spam. Google states plainly that “buying and selling links is a normal part of the economy of the web,” and that a paid link is not a violation “as long as they are qualified with a rel="nofollow" or rel="sponsored" attribute.” The line is not whether money moved. It is whether the paid link is allowed to pass ranking credit.
Cross that line and the examples in Google’s spam policies start to fit the page: advertorials and paid guest posts whose links pass ranking signal, optimized anchor text sold inside articles, forum comments stuffed with promotional links. A publisher who runs sponsored posts without the sponsored attribute is, by that definition, selling ranking signal, whether or not anyone framed it that way.
For the blogger from the opening, the fix is small and the exposure was real. One rel="sponsored" on the brand link turns a policy problem back into an ordinary advertisement.
When links change hands, or you are not sure #
Two situations resist a clean rule. The first is a link that ages into a paid relationship: an old editorial mention of a tool you later signed an affiliate deal with. The honest move is to re-qualify it when the relationship changes, because the attribute describes the link’s current status, not its origin. The second is genuine uncertainty. When a link does not obviously fit sponsored or ugc but you are not comfortable vouching for it, nofollow is the designed catch-all, and Google’s guidance is that existing nofollow links require no conversion. Reaching for nofollow when unsure is not a workaround. It is what the value is for.
What you cannot do is nothing. A naked outbound link on a page that took money, or a comments section that auto-follows every URL a stranger drops, is the exact state Google’s link policies were written against.
The three attributes will not rescue a weak linking strategy, and they were never built to. What they do is narrow and worth doing well: they let you tell Google, one link at a time, which of your outbound links carry your endorsement and which do not. That is a small honesty, filed one tag at a time, and on a site that publishes ads or hosts a comments section it is the difference between an ordinary web page and one a competitor can report. The blogger who now marks her paid links is not gaming anything. She has just stopped vouching for links she was paid to place.