Guides  ·  1,900 · KD 8 searches a month

Do nofollow links help SEO

You paid for a link, it came back nofollow, and somebody told you it was worthless. That was true until March 2019, and it has not been true since.

Part 1

It became a hint, not an instruction

In September 2019 Google announced that nofollow would be treated as a hint for ranking purposes, effective March 2020. Hint means they may use the link, and they will not tell you when. So the honest answer is that a nofollow link is worth something unknown and greater than zero.

Do this: Stop rejecting placements because the link is nofollow. Judge the placement by whether real people will see it.
Part 2

The traffic was always the point

A link on a page that people read sends you visitors, and those visitors are worth more than the ranking signal you are arguing about. A dofollow link on a page nobody visits sends you nothing at all.

Do this: Ask for the page’s traffic before you ask about the link attribute. If they cannot tell you, the link is decoration.
Part 3

The new attributes changed the question

Alongside nofollow there is now rel="sponsored" for paid links and rel="ugc" for comments and forums. Using sponsored on a paid link is what Google asks for, and refusing to mark a paid link is what gets sites into trouble.

Do this: If you paid for it, mark it sponsored. The risk of not marking it is much larger than the value of the signal.
We counted

We counted 497 outbound links on reference sites. Not one used sponsored or ugc.

Everyone quotes the 2019 announcement. Nobody checks what the web actually does with it. So we pulled six reference pages, parsed every outbound anchor, and read the rel attribute on each one.

Page Outbound nofollow Share
Wikipedia · PageRank23221492%
Wikipedia · SEO15113287%
Wikipedia · Nofollow584272%
Hacker News front page31825%
MDN · rel attribute2400%
Total49739679%

Three things fall out of that table, and the third is the one worth your attention.

Nofollow is the default on user-editable pages. Wikipedia runs between 72% and 92%. That is policy, not judgement about any individual link, and it is why a Wikipedia citation was never going to pass what people hoped it passed.

Editorial pages barely use it. MDN links out 24 times and marks none of them. When a human editor chooses each link, the attribute stops being interesting.

Across all 497 links, sponsored appeared zero times and ugc appeared zero times. Both attributes have existed since 2019. In this sample the web has not adopted either one.

What it means for you: the distinction between nofollow, sponsored and ugc matters far more on the pages you publish than on the pages that link to you. Almost nobody is marking their links precisely, so Google is reading context regardless.

Method: six pages fetched 2026-08-23, every anchor with an absolute href to another domain parsed, rel attribute read. Four further pages were attempted and excluded: two returned an error to a non browser request, two rendered their links with a script. A sample of six pages is small and skewed toward reference material, which is exactly where nofollow policy is strictest.

Part 4

Where the effort should actually go

A link is one signal. A page nobody can crawl, an intent you did not match, or a title nobody clicks will each cost you more than the difference between nofollow and dofollow ever will.

Do this: Run the free page check first. Fix what it finds before you spend another hour on link attributes.

This is teaching material and our own reading of standard practice, not advice for your specific site. Check anything important with your own specialist.

Free consultation

Want us to look at your page? It costs nothing.

Run the free check, then tell us what it found. We reply with what to fix first and why. No call, no pitch.

Run the free check →

If a word here is unfamiliar it is defined in the SEO glossary. Two public penalties and one problem we found on our own network are pulled apart in the case studies, and the longer pieces live in the library.