Google says traffic through a link is not used to value it. What Mueller said, what the reasonable surfer patent and an Ahrefs study show, and what to check.
No, not as far as Google has said. Google's John Mueller has said Google does not use the traffic that goes through a link to decide what that link is worth. A link nobody clicks can still pass value, and a link that sends plenty of visitors does not pass more because of those visits. Clicks matter for a different reason: they bring people to your site. Here is what Google has said, what the reasonable surfer patent and a large Ahrefs study show, and how to judge the links you already have.
What Google has said about clicks and link value
The question came up in a Google Search Central SEO hangout recorded in September 2021. Mueller was asked whether a link passes more value when it sends more traffic. As reported by Search Engine Journal, he said Google does not use "traffic through a link when trying to evaluate how a link should be valued". He added that, as far as he knew, Google does not look at "the probability that someone will click on a link with regards to how we should value it".
His reason was that links are sometimes just references. A reader can see a citation and move on without clicking it, and he said he did not think Google would take that into account when valuing the link.
He then pointed to how PageRank works. Each page has some value, and it passes a fraction of that value on through its links. In his words, if a page is seen as very high value, "the links from there will be treated with a little bit more weight compared to some random page on the internet". So in his framing, what a link is worth depends on the value of the page it sits on, not on how many people click it. He also said Google handles PageRank "a bit different nowadays".
Two limits apply. This was a spoken answer in a video hangout, not a documentation page, and Mueller hedged the second part himself. It is still a direct answer to this exact question.
What the reasonable surfer patent says
Arguments about links from pages with traffic have run since SEO practitioners found Google patent US 7,716,225 B1 titled "Ranking documents based on user behavior and/or feature data", filed in 2004 and granted in 2010. Ahrefs calls it Google's Reasonable Surfer patent. You can read it on Google Patents.
Its abstract describes a model built from features of a link along with user behavior data about how people move through links, and a rank assigned to a document based on that model. The patent's idea is that a surfer will follow some links on a page with higher probability than others. It lists link features such as the font size of the anchor text and where the link sits: in running text, in a list, above or below the first screen, in a footer or in a sidebar. It gives "Terms of Service" links and banner ads as examples of links people are unlikely to follow.
The detail that matters for this question is how a single link gets its weight. The patent says: "The weight of a link may be a function of the rules applicable to the feature data associated with the link." It also says a link's weight "may reflect the probability that the link will be selected". Click data does feed the model: the link data it describes "might include how often the link was selected", and the model may learn rules for a single document from user behavior data. As described, a link's weight is the model's estimate of how likely that link is to be selected, worked out from its features.
A patent cannot tell you whether Google uses the system today, or in what form. Google Patents lists this one as expired.
What court filings add
In the US v. Google antitrust case, the plaintiffs' post-trial brief, published by the Justice Department, calls Navboost "one of the most powerful ranking components" at Google and says it "memorizes all clicks for all queries received in the prior 13 months". That is about clicks on search results, not clicks on links between websites. It does not show that click data is used to value backlinks, and Mueller's answer above says Google does not use traffic through a link to value it.
What the data shows about links from pages with traffic
A related question is whether links from pages that get search traffic are worth more. Tim Soulo of Ahrefs tested this in a study of links with traffic. Ahrefs took 44,589 non-branded keywords with monthly search volumes between 2,000 and 5,000, pulled the top 10 pages for each, and checked how well different link counts lined up with ranking position.
What it found:
- Links from pages with traffic did not beat plain link counts. Soulo had expected links from pages with traffic to correlate more strongly than the total number of links, and they did not.
- The amount of traffic mattered more than the count of pages with traffic. The sum of organic traffic to linking pages correlated better than the number of linking pages with traffic, which Soulo read as a sign that links from more powerful pages carry more weight.
- At the site level, traffic made little difference. Ahrefs concluded that "the amount of organic traffic to a referring domain has little impact on the value of a link".
- Links with traffic were rare among top pages. In only 20 of the results pages studied did every top page have at least one link from a page with traffic, and in about one in five, none of the top pages had any.
The study has limits that Ahrefs states itself. It measured estimated organic search traffic to the linking pages, not referral traffic and not clicks on the links. It is a correlation study. Soulo's suggestion from the data was to prioritize links from pages with a high URL Rating and from pages on sites with lots of referring domains. That is in line with Mueller's remark that links from a page seen as very high value get a little more weight.
Claims to treat with care
Three related claims come up in this debate, and none of the sources above support them: that Chrome data feeds link valuation, that links on pages without traffic are ignored, and that a link profile with real clicks is safe from spam updates. Google's own spam policies describe link spam by how links were created, such as "Buying or selling links for ranking purposes" or placing links with optimized anchor text in articles and guest posts distributed on other sites. None of the link spam examples turn on whether anyone clicked.
A test you can run yourself
Clicks still have a use that does not depend on rankings. A link that sends visitors is worth having for the visitors, and you can measure that. This is our method, not a Google or Ahrefs one.
For every link you earn or place, record the date it went live, the page it sits on, and where it sits on that page. After a month, open your analytics and look at referral sessions from that page. Group the links into those that sent visitors and those that sent none. Then compare, for the two groups, what you can see over the same period in Search Console for the target page: impressions and clicks for its main queries.
You will get a small, honest dataset about your own site. If the two groups look the same, you have learned that on your site, clicks alone did not separate them. If they differ, remember that the two groups may also differ in the strength of the pages they sit on, so the cause is not settled. Either way you will know which links bring readers.
What to do with a link nobody clicks
Do not remove a link because it sends no visitors. Look at the page first. Is it live, is it indexed, is the link followed? SerpCrest's bulk DoFollow checker answers those three questions for a list of URLs. Then look at the linking page itself: whether it is a page that other sites link to. Mueller pointed to the value of the linking page, and in the Ahrefs study URL Rating was the strongest correlating factor. Our post on what changes after you build links covers a Google patent that describes rankings dipping after new links, and the piece on nofollow links and AI search covers why a link with no ranking credit can still matter.
Judge a link by the page it sits on, not by its clicks. Place links where a reader would plausibly click them anyway, because the visitors are a benefit you can measure.
