Dofollow vs Nofollow vs Sponsored vs UGC Links: What Actually Moves Rankings in 2026

Every backlink you earn or place carries an invisible tag. Most people never look at it. That tag, the rel attribute sitting inside the anchor's HTML, is the difference between a link Google treats as a vote and a link Google treats as a mention, and the line between those two is blurrier than either camp likes to admit. This note lines up the four attribute states you will actually run into, dofollow (the default, no attribute at all), nofollow, sponsored and UGC, against what Google's own documentation says each one does and what a handful of named practitioner tests say actually happens once a link goes live.

Our usual disclaimer applies. We sell nothing here, no placements and no packages, and everything below is either quoted from Google's own Search Central documentation, drawn from Google's official announcement of the sponsored and ugc attributes, or credited to a named practitioner test we link to directly. Where a claim rests on one team's single test rather than a controlled study across many sites, we say so plainly.

What each rel value actually tells Google

Dofollow is not a real attribute value. It is shorthand the industry invented for "no rel attribute at all," which is the default state of every link and the only one that passes full ranking credit by design. The other three are real values defined in Google's own developer documentation. rel="sponsored" marks links that exist because of a payment, sponsorship or other compensation arrangement: paid placements, affiliate links, paid guest posts. rel="ugc" is recommended for links inside user generated content, comments and forum posts specifically. rel="nofollow" is the catch-all: Google's documentation says to use it "when other values don't apply, and you'd rather Google not associate your site with, or crawl the linked page from, your site." The same documentation notes that nofollow was the attribute previously recommended for paid links and is still an acceptable substitute for rel="sponsored", though sponsored is now Google's preferred tag for that case.

From directive to hint: the shift that changed the math

For fourteen years after nofollow's introduction in 2005, a joint effort by Google, Yahoo and Microsoft to fight blog comment spam, Google treated the attribute as absolute: a nofollow link passed no ranking credit, full stop. That changed on September 10, 2019, when Google's Search Central team introduced rel="sponsored" and rel="ugc" alongside a structural change to how all three attributes function. The announcement reclassified nofollow from a rule Google always obeyed into a hint Google could choose to follow, and that hint treatment extended to crawling and indexing decisions on March 1, 2020. Google's current documentation, last updated in December 2025, still describes the same model: links carrying any of the three attributes "will generally not be followed," but the pages they point to "may still be crawled" through other paths such as sitemaps or links from other sites, and Google keeps the option to use the link as a ranking signal when its systems judge that useful.

Does a nofollow link actually move rankings?

This is where the industry argument actually lives, and the honest answer is sometimes, on Google's terms, not yours. The most cited practitioner test comes from Joy Hawkins at Sterling Sky, a Canadian local SEO agency. In a test first published in November 2024 and still cited through 2026, Hawkins posted a nofollow comment link on a high-traffic, well-ranking page on Local Search Forum, pointing to a page on Sterling Sky's own site. The next day, a YouTube video embedded on the linked page began appearing in search results for the targeted keyword, on a page that had not ranked for it before. Hawkins reports repeating the pattern with nofollow links from other high-traffic pages afterward, with consistent results. That is one agency's repeated but informal test, not a controlled study across many domains, and we are flagging it as exactly that, the same way we flag our own single-property experiments elsewhere on this site. What it does line up with is Google's stated model: a hint system that can weight a nofollowed link more heavily when it sits on a page that itself ranks and pulls real traffic, and closer to zero when it sits on a thin, untrusted page. The attribute alone is no longer the whole story. The page carrying the link is now part of it.

Building a profile that looks earned, not manufactured

Google's spam policies require that any link acquired through payment, product exchange or another compensation arrangement carry rel="sponsored" or, at minimum, rel="nofollow", and that compliance floor has not moved. Outside of it, the more useful question is mix. Moz's published guidance on link attributes, echoed by most major SEO toolmakers, is that a backlink profile made entirely of dofollow links reads as unnatural, because real, widely linked sites accumulate a blend on their own: dofollow links from editorial coverage, nofollow mentions from large publishers who nofollow everything by policy, UGC links from forum threads and reviews, and the occasional sponsored disclosure. A backlink profile that is 100 percent dofollow is itself a signal. It is just not the one anyone building it wants to send.

What this means for the next link you're offered

1. Check the attribute, then check the page. A nofollow link on a page that ranks and gets visited is worth more than the attribute alone suggests.
2. Keep compliance separate from strategy. Paid is sponsored or nofollow no matter what you hope it does for rankings.
3. Do not reject a placement on reflex just because it is nofollow. High-traffic nofollow real estate can outperform a dofollow link on a page nobody reads.
4. Watch your own mix. A backlink profile that is entirely dofollow is a pattern worth investigating, not a result worth celebrating.

None of this argues for chasing nofollow links over dofollow ones. Dofollow remains the only state that passes ranking credit by default, and it should stay the priority whenever you have a real choice between the two. It is an argument for not discarding a strong nofollow opportunity on reflex, and for treating the linking page, not just its rel attribute, as the thing you are actually buying. We are now tracking attribute type against twelve-month ranking movement across the placements we monitor, the same way we track survival rates elsewhere on this site, and we will publish what we find, including if it turns out to be nothing. If you want that when it lands, along with one dataset and one thing to stop paying for in every issue, the Loft Brief goes out twice a month and costs exactly what independent data should cost you to read: nothing.

Frequently asked questions