Nofollow vs Dofollow: Why the Old Binary Stopped Working

TL;DR: The nofollow vs dofollow question is the wrong frame. The real question is whether the link is an editorial endorsement, because Google treats rel attributes as hints while still judging the source page, anchor fit, link context, and why the link exists.

Most people search “nofollow vs dofollow” expecting a clean switch: one link passes SEO value, the other does not. Tidy. Also outdated.

At mindnow, vadimkravcenko.com, and seojuice.com, I’ve seen the same mistake from three angles — developers add nofollow like it is a firewall, marketers chase any “dofollow” backlink like the attribute alone makes it valuable, and editors nofollow every outbound link because someone told them link equity leaks out of the page.

There is no rel="dofollow" attribute. A normal link is just a normal link. nofollow, sponsored, and ugc are labels that describe the relationship behind the link. Treat them like disclosures, not magic SEO valves.

The short answer: what nofollow and dofollow mean now

A normal HTML link looks like this:

<a href="https://example.com">Normal link</a>

SEOs call that a “dofollow” link, but the browser and HTML spec do not. It is just an <a> tag without a restrictive rel value. In SEO language, it means you are linking normally and are not asking search engines to treat the link differently.

A nofollow link looks like this:

<a href="https://example.com" rel="nofollow">Nofollow link</a>

nofollow tells Google you do not want to give the same endorsement you would give with a normal editorial link. Google may still look at it, but it is treated as a hint (a hint, not a command).

The useful rule is simple. Use a normal link when you are willing to endorse the destination. Use nofollow when you need to reference something without endorsing it. Use sponsored when money, products, or compensation are involved. Use ugc when users added the link instead of your editorial team.

I got this wrong before I audited enough real sites (I got this wrong for years). I treated the attribute like the whole link. It is only one field in a much larger evaluation.

The old model was simple. The new model is messier.

Nofollow started as a spam control tool. Blog comments were getting hammered with links, and site owners needed a way to say, “I did not vouch for this.” The early mental model was easy to teach: normal links pass PageRank, nofollow links do not.

That model stuck because it was convenient. It made link reports easy. It made sales pitches easy. It gave junior SEOs a binary filter in Ahrefs, Semrush, and every other backlink tool.

Google changed the frame in 2019.

All the link attributes — sponsored, UGC and nofollow — are treated as hints about which links to consider or exclude within Search.

Danny Sullivan and Gary Illyes wrote that in Google’s “Evolving nofollow” announcement. The sentence matters because it breaks the locked-gate idea. A hint can be used. A hint can be ignored. A hint is not the same as a crawling block.

John Mueller made the same distinction when comparing nofollow with robots.txt:

So it's not a 100% directive like robots.txt, where you say these are never going to be crawled. But it does tell us that we don't need to focus on them as much.

Robots.txt can block crawling. A nofollow attribute sits on an individual link and gives Google a relationship signal. Those mechanisms solve different problems.

Google also explained why it moved to hints:

By shifting to a hint model, we no longer lose this important information, while still allowing site owners to indicate that some links shouldn't be given the weight of a first-party endorsement.

The key phrase is “important information.” Links contain more than PageRank. They contain anchors, context, discovery paths, source relationships, and patterns.

Why a nofollow link can still matter

A nofollow link can matter without becoming a clean PageRank vote. That distinction keeps the argument honest.

Google may still use nofollow as a hint. The anchor text and surrounding content may help Google understand what the destination is about. The link may help discovery. The referral traffic may turn into sales, newsletter subscribers, branded searches, or future links from people who found the page through that citation.

Google said this directly in the same 2019 post:

Links contain valuable information that can help us improve search, such as how the words within links describe content they point at.

A nofollow link from a major publication may not behave like a normal editorial link from a smaller industry blog. Fine. It can still send qualified visitors, put the brand in front of writers, and trigger later citations that are normal editorial links. Attention moves like that — none of it is magic.

Think of a SaaS company mentioned in a nofollowed review on a large software comparison site. The link may be tagged. The traffic still arrives. Some visitors search the brand later. One consultant writes a roundup and links normally. A founder mentions the tool in a podcast note. The first link was not the whole outcome; it was the start of a path.

Bad SEOs turn that into “nofollow links pass ranking power.” I would not say that. A better sentence is: nofollow links can participate in discovery, context, demand, and downstream citation. In some cases, those effects matter more than the raw attribute.

Why a dofollow link can still be worthless

A normal link only means the link lacks nofollow, sponsored, or ugc — it does not make the source page trustworthy, the anchor relevant, the page indexed, or the placement editorial.

This is the part most beginner guides underplay. A normal link from a weak page can be close to zero. A normal link from a spammed, irrelevant page can become a liability. Attribute status cannot rescue a bad source — quality still does the heavy lifting.

The 2024 Google Content Warehouse API leak did not give SEOs a public ranking formula. Be careful with anyone who says it did. Still, analyses from Mike King and Rand Fishkin reinforced something link builders should already know: Google’s link systems appear to care about source quality, tiers, anchor matching, and link context.

sourceType…is to record the quality of the anchor's source page and is correlated with but not identical to the index tier of the source page. Effectively, this is saying the higher the tier, the more valuable the link.

That is Mike King’s reading of the leaked documentation in his iPullRank analysis. The important idea is source quality, not the word “dofollow.”

King also called out anchor mismatch:

Anchor Mismatch – When the link does not match the target site it's linking to, the link is demoted on the calculations.

Rand Fishkin made the broader point in his SparkToro analysis:

PageRank still appears to have a place in search indexing and rankings, but it's almost certainly evolved from the original 1998 paper.

That is the adult version of link SEO. Google still cares about links, but not in the childish “dofollow good, nofollow bad” way. A normal link from a real editorial page, with a sensible anchor and strong topical fit, is different from a normal link buried in expired-domain sludge.

The decision rule: choose the rel attribute by relationship, not by SEO wish

The rel attribute should describe why the link exists. Do not choose it because you want more ranking value. Do not remove it because a sponsor asked for “SEO juice.” Do not add it to every outbound link because someone in 2012 scared you.

Glenn Gabe’s framing is the cleanest version I’ve seen:

The nofollow attribute is a way to tell Google that the link did not earn the vote. It was paid for, sponsored, added as user-generated content, etc.

Here is the practical model.

Multiple values are valid. rel="ugc sponsored" is a real thing. So is rel="ugc nofollow". The values overlap in effect, but they do not mean the same thing. They describe different origins and relationships.

The decision gets easier when you stop asking, “Will this pass value?” Ask, “Did this link earn a first-party editorial vote?” If yes, leave it normal. If no, label the reason.

Nofollow vs sponsored vs ugc

nofollow means you are not endorsing the linked page, or you do not want your site’s reputation associated with it. Use it for references you need to include but do not trust.

sponsored means the link exists because of money, product, service, affiliate revenue, or another commercial relationship. Paid links should be disclosed with sponsored (in 2026, this is no longer a technicality).

ugc means user-generated content. Comments, forum posts, profile bios, community submissions, marketplace listings, and public Q&A answers often fall here. The site may host the link, but the editorial team did not place it.

Do not treat the three attributes as interchangeable. A paid link is not merely “untrusted.” A user link is not automatically an ad. An untrusted editorial reference is not user-generated content.

Messy cases exist. A community member might post a paid promo link in a forum. Use rel="ugc sponsored". A commenter might link to a source you do not want to vouch for. Use rel="ugc nofollow". The attribute should describe reality, not your preferred SEO outcome.

Should you nofollow internal links? Usually no.

Internal nofollow is usually a sign that site architecture is doing work it should not be doing. If a page should not be crawled, handle that with robots rules, canonical strategy, authentication, or better architecture. If a page should not rank, ask why it is public and linked from important pages.

Old PageRank sculpting created a bad habit. SEOs tried to nofollow internal links to force more equity through selected paths. Google changed how it handled that years ago, and the habit never really deserved the confidence people gave it.

Normal internal links help Google understand your site. They show hierarchy, related pages, hubs, supporting articles, and commercial paths. At seojuice.com, we care about internal linking because links create meaning inside a site, not just crawl paths.

Do not nofollow normal links from blog posts to related articles. Do not nofollow breadcrumbs. Do not nofollow pagination by default. Do not nofollow contextual internal links just because you want one money page to get all the attention.

If anchor text is messy, fix the anchor text. If pages are orphaned, run an internal link audit. If crawl waste is real, work on crawl budget, templates, canonicals, and indexation rules.

A rare exception might be untrusted user-added internal links in a community product. Even then, I would ask whether permissions, moderation, or URL handling should solve the problem instead (and yes, I still see this).

Backlink audits: stop sorting only by dofollow

Filtering a backlink export by “dofollow only” is a lazy first pass, not an audit. It may help you reduce a spreadsheet, but it also hides useful links and overvalues junk.

A serious link review asks better questions. Is the linking page indexed? Is it relevant? Is the link stable? Is it visible to users? Does the anchor match the target page? Is the link surrounded by real content? Does the site appear in a pattern of manufactured placements?

A nofollow citation from a Wirecutter-style review may not pass classical PageRank, but the qualified traffic, journalist exposure, and downstream mentions can beat fifty normal links from expired-domain blog networks.

Use this order when reviewing links:

1. Source page quality.
2. Topical fit.
3. Anchor fit.
4. Link placement.
5. Traffic or visibility.
6. Pattern risk.
7. Rel attribute.

The order is the point. Rel comes last. A normal link on a bad page is still a bad link. A nofollow link on a trusted, visible page can still be worth understanding.

Seojuice.com looks at link context because internal and external links work through meaning, not only graph math. A link is a sentence fragment with a source, a target, and intent.

Common myths about nofollow vs dofollow

“Nofollow links never help SEO”

False. They may not pass equity like normal editorial links, but they can still help discovery, context, referral demand, and downstream citation. Google moved to a hint model because link data still contains useful information.

“Dofollow links are always good”

False. Bad links can be ignored, discounted, or become part of an unnatural pattern. A normal attribute cannot make a weak source strong.

“You should nofollow every outbound link”

False. Linking to good sources helps users and clarifies your page’s context. Fear-based outbound linking makes pages worse. If you cite a trusted source, a normal link is appropriate.

“Sponsored links should just be nofollow”

Outdated. Google created sponsored so site owners can describe paid relationships more clearly. If the link is paid, use sponsored. If it is also user-generated, combine values.

Final rule of thumb

If the link earned an editorial vote, leave it normal. If it was paid, mark it sponsored. If a user added it, mark it ugc. If you need to reference something without endorsing it, use nofollow.

The attribute matters, but it is not the link. The link is the source, the context, the anchor, the page, and the reason it exists. Dofollow versus nofollow is only one part of that story.

FAQ

Is dofollow a real HTML attribute?

No. “Dofollow” is SEO shorthand for a normal link without nofollow, sponsored, or ugc. Adding rel="dofollow" is unnecessary.

Can Google ignore nofollow?

Google says nofollow is a hint. That means Google can choose to consider or exclude the link in different systems. It should not be treated like robots.txt.

Should affiliate links be nofollow or sponsored?

Use rel="sponsored" for affiliate links. If your CMS only gives you one field, sponsored is the clearest label for the commercial relationship.

Are nofollow backlinks worth reporting?

Yes, if the source is relevant, visible, and capable of sending traffic or creating later citations. Do not report them as guaranteed ranking equity. Report what they actually did.

Should I ask sites to remove nofollow from links to me?

Usually no. If the link is paid, user-generated, or non-editorial, the attribute belongs there. If the site genuinely cited you as a source and added nofollow by policy, you can ask once. Do not build a strategy around begging for attribute changes.

Fix the relationship, not just the attribute

SEOJuice surfaces the relationships between your pages so you stop arguing about a single rel attribute and start fixing the architecture under it. If your internal links, anchors, and content paths need a cleaner model, start with an internal link audit.