Yes. Anchor text is still a Google ranking factor in the sense that matters: Google uses it to understand what a linked page is about, and that function never went away after Penguin. What Penguin changed was how Google treats manipulative linking, not whether the words inside a link carry meaning. Google's own link guidance, last updated in December 2025, says good anchor text is descriptive and relevant to the page it sits on and the page it points to. That is years after Penguin, and it reads nothing like a retired signal.
Call the evidence here strongly supported, not confirmed outright. Google has told us anchor text helps it understand a destination page. Nobody outside Google has told us how much that single signal is worth next to everything else in the ranking system, and nobody has published a safe exact-match percentage. So the honest answer has two parts: yes, anchor text still works the way you'd expect, and no, there is no magic ratio that protects you or trips an alarm. Treat anything that promises a precise number here with real skepticism.
Why Penguin made anchor text controversial
The anchor text ranking factor question only became a question because of what Google did in 2012. On April 24, 2012, Matt Cutts announced an update aimed at webspam, naming keyword stuffing and link schemes as the kind of manipulation it would push down in results. Google showed examples of unrelated outgoing links stuffed into low-quality, spun text. It did not publish the specific signals it used, because disclosing them would have handed a roadmap to the people trying to game the results.
That gap between "Google acted against manipulative linking" and "here is exactly what changed about anchor text" is where a lot of confusion took root. People watched sites with aggressive exact-match anchors lose rankings and reasonably concluded that anchor text itself had become dangerous. Ahrefs, writing about this period, describes manipulative anchor text as one of Penguin's main targets and reports that sites leaning too hard on exact-match anchors lost ground. That is a fair reading of what happened to those specific sites. It is not the same as Google stating a before-and-after number for how much anchor text counts.
The next chapter of Penguin anchor text history matters just as much as the first. Penguin anchor text enforcement changed again in September 2016, when Google folded Penguin into its core algorithm. Two things shifted. First, the list of sites affected started refreshing in real time instead of on a periodic schedule, with changes kicking in after Google recrawled and reindexed the relevant pages. Second, Google said Penguin began devaluing spam signals with more granularity, adjusting rankings based on those signals rather than dragging down an entire site at once. Google's own wording: Penguin now devalues spam by adjusting ranking based on spam signals, rather than affecting ranking of the whole site. That is a real change in how enforcement works. It is not a disclosed formula, and "real time" still means after a crawl happens, not the instant you edit a page.
None of this tells you a safe percentage. The 2012 and 2016 announcements describe a shift in how Google treats spam and how fast it reacts. They do not hand you a published exact-match ceiling, a universal amount by which anchor weighting dropped, or proof that every keyword-rich anchor became unusable. Keep that distinction in mind, because a lot of advice floating around collapses it.
The evidence that anchor text still matters
The strongest evidence for anchor text as a ranking factor is current, not historical. Google's Search Central link guidance, last updated December 10, 2025, says anchor text tells people and Google something about the page being linked to, and that better anchor text helps Google understand what that page is about. The guidance tells writers to write naturally, avoid cramming every related keyword into the anchor, skip vague phrasing when something specific would help, and pay attention to the sentence around the link. It also says to use internal anchor text so people and Google can understand the site and find relevant pages.
That is not the language of a retired signal. It postdates both Penguin announcements by years, and it directly contradicts the idea that Google stopped caring what the clickable words say.
Internal linking gets its own confirmation. Reporting on comments from Google's John Mueller in a Webmaster Central discussion, Search Engine Journal described how internal linking helps Google find pages and gather context, some of it from anchor text and some from where a page sits in the site's structure. In a specific exchange about linking to related pages, Mueller said reasonable anchor text was fine and that the publisher did not need to invent elaborate alternative phrasing. He also said a particular proposed tweak was not something the publisher would see a visible effect from in Search. Read that carefully: it means a small wording change on one internal link might not move the needle, not that anchor text does nothing. Those are different claims, and mixing them up is how a reasonable comment turns into an overstated one.
Put the two together and you get a consistent picture. Google uses anchor text, inside and outside your site, to build an understanding of what a page covers. That understanding is real. What it is not is a dial you can turn with precision, because Google has never published the weight it carries next to everything else feeding a ranking.
What the research actually shows about anchor ratios
Here is where a lot of advice overreaches. Ahrefs published a widely cited anchor-text study, examining the top 20 results for 19,840 keywords, around 384,614 pages total. For exact-match anchors, the correlation with ranking position came out weak: 0.1436 using average exact-match percentage and 0.1869 using the median. Ahrefs itself called the relationship weak and noted the median exact-match share was zero at every ranking position in its data. Surrounding-text matches, where the target phrase appears near a link rather than inside it, scored even lower: 0.0640 and 0.0205, which Ahrefs described as negligible.
A couple of things matter about that data before you build a strategy on it. First, it is a correlation, gathered by observing existing pages, not a controlled experiment where anchor text was the only thing changed. Other differences between those pages, including link quality and internal linking patterns, could easily be doing some of the work the correlation seems to credit to anchor wording. Second, the study is from April 2020. Google's current guidance is from December 2025. Treating a five-year-old correlation as a live measurement of today's algorithm is a mistake, however often the number gets repeated.
What the research does rule out is the simple story that stacking exact-match anchors is a reliable path to higher rankings. A correlation around 0.14 to 0.19 is not nothing, but it is nowhere near strong enough to support "more exact match equals better rankings" as a rule you can act on. What it does not rule out is that anchor text plays a real, if modest and hard to isolate, role in how Google reads a page. Weak correlation and zero effect are not the same finding, and the honest middle ground between them is uncomfortable for anyone who wants a clean number to chase.
Resist the urge to turn 0.1436 into a weighting percentage. A correlation coefficient describes an association across a large sample. It does not tell you that anchor text contributes 14 percent of anything, and nobody who has actually seen Google's ranking code has said otherwise.
What an over optimized anchor text penalty actually means
The phrase over optimized anchor text penalty gets used as if it names one thing. It actually names three different mechanisms, and mixing them up leads to bad decisions.
The first is a policy violation. Google's spam policies define link spam as links created mainly to manipulate rankings, and the examples specifically include optimized anchor text in guest posts, articles, or press releases distributed on other sites, plus keyword-rich links buried in widgets or footers distributed across many pages. That is about the practice and its purpose, not about hitting a particular percentage. A descriptive, exact-match anchor placed by an editor who genuinely found your page useful is not a violation. A manufactured placement with a branded anchor instead of an exact-match one does not become compliant just because the anchor looks tamer.
The second is algorithmic devaluation, which is what Penguin actually does since 2016: adjusting rankings based on detected spam signals rather than handing out a visible penalty. This can happen quietly, with no notice, no manual review, and no entry in your Search Console reports.
The third is a manual action, which is different again. Google's manual actions guidance describes this as a human reviewer determining that specific pages or a whole site violate the spam policies, which then shows up as a flagged item in Search Console with a path to fix the issue and request reconsideration. A concentration of exact-match anchors in a backlink report is an observation. On its own, it tells you nothing about which of these three situations, if any, is actually happening to you.
Disavowing links is sometimes treated as a cleanup step for an anchor-text ratio that looks aggressive. Google's own disavow guidance says most sites never need this tool, because Google is generally able to judge which links to trust on its own, and it warns that misusing disavow can hurt your search performance. Reserve it for a real volume of spammy or artificial links tied to an actual or likely manual action, not for routine ratio maintenance.
What anchor-text ratios should you use
There is no universal safe percentage in anything Google has published. Ahrefs reports that an industry convention of roughly 1 to 5 percent exact-match anchors has circulated for years, but it presents that explicitly as an industry habit, not a Google-approved ceiling, and ultimately advises against engineering your ratios to hit it. Separately, in comments reported by Barry Schwartz in October 2018, John Mueller said anchor-text percentages and similar counts were not indicative of quality on their own. That is a direct statement against treating an isolated ratio as a safety score, from someone at Google, not an inference from a third party's dataset.
A practical policy beats a fabricated formula here. For editorial links you write, pick the clearest accurate description of the destination, in context, and do not rotate phrasing mechanically to hit a quota. When another publisher independently links to you with an anchor that happens to be exact match, that is a good sign, not a problem to fix. If you run a campaign that places the same commercial anchor across many sites, the thing to examine is the placement strategy itself: editorial independence, how the links were obtained, and whether paid or sponsored links carry the right rel attributes. Switching those links to a branded anchor instead of properly qualifying them does not solve anything.
Ratios also need a denominator before they mean anything. A page with 10 exact-match links out of 100 total links sits at 10 percent. Add 900 unrelated branded links and the share drops to 1 percent, even though none of the original 10 links changed. That arithmetic shows why a ratio by itself, without knowing whether you are counting links, referring pages, or referring domains, and without knowing the target query, cannot diagnose anything on its own. Use anchor-text data to investigate a profile you already have reason to be curious about. Do not set an acquisition target like "keep exact match under 3 percent" and call that evidence-based risk management, because no published source supports that specific number as a safety line.
What would change this verdict
The honest answer to the anchor text ranking factor question will stay qualified until Google says more. A future, more specific disclosure from Google about anchor-text weighting would change how precisely this can be answered. So would a well-designed experiment that isolates anchor wording from every other variable, something the existing correlation studies do not do. Until either of those shows up, the honest position stays the same: anchor text is a real relevance signal Google has confirmed it uses, and no one, including the vendors who publish ratio guidelines, has a verified number for how much of your ranking it decides.
If your team is spending hours auditing anchor-text spreadsheets every time a new backlink report comes in, that time is usually better spent on the editorial side of linking: writing clear, accurate anchors as a habit, and using your own internal links to describe pages the way you'd want Google and your readers to read them. DeepSmith's writing pipeline builds internal linking into article generation itself, pulling from a site's own content map rather than leaving it to a manual pass after the draft is done, which is the step most teams admit they skip when they're behind schedule.
