Anchor text is one of the oldest ranking signals in Google's algorithm - and one of the most abused. The clickable words that form a hyperlink communicate far more than a destination. They tell Google what the linked page is about, how relevant the source page considers it, and - when viewed across hundreds of links - whether a site is earning links naturally or manufacturing them. Getting anchor text right means understanding all three dimensions at once.
Why Anchor Text Matters for SEO
When Google's crawlers follow a link, they read the anchor text as a relevance signal for the page being linked to. A link pointing to a page about technical SEO audits with the anchor text "technical SEO audit guide" tells Google that the destination page covers that topic - and that the source page considers it worth referencing under that label. Aggregate enough of those signals from enough different sources, and the linked page builds topical authority for that phrase.
This mechanism is why anchor text became a primary target for manipulation. If the signal is strong, engineers will try to engineer it. For a long time, they succeeded. Then Google adapted.
The deeper point is that anchor text is not just a ranking lever. It is a user experience signal. Descriptive anchors help readers understand where a link leads before they click, which increases click-through rate and reduces bounce from unexpected destinations. Both factors feed back into how Google evaluates page quality.
The Five Anchor Text Types
Anchor text falls into five broad categories, each carrying different risk and relevance profiles:
Exact-match: The anchor text matches the target keyword precisely. Example: a link to a page optimised for "WordPress schema plugin" using the anchor "WordPress schema plugin." High relevance signal, high manipulation risk.
Partial-match: The anchor includes the target keyword alongside other words. Example: "the best WordPress schema plugin for developers." Carries most of the topical signal with far less footprint risk.
Branded: The anchor uses the brand name - "Signocore," "Moz," "Ahrefs." These anchors build brand authority and are almost impossible to over-optimise because no competitor is trying to rank for your brand name.
Naked URL: The URL itself is used as anchor text - "https://example.com/guide." Common in citations, press mentions, and content scrapers. Neutral topical signal, but adds diversity to a link profile.
Generic: Non-descriptive phrases like "click here," "read more," or "this article." Carry almost no topical signal and contribute nothing to keyword relevance. Acceptable in small quantities; dominant use is a wasted opportunity.
A healthy site's inbound link profile contains all five types. The exact proportions vary by industry and link acquisition method, but no single category - especially exact-match - should dominate.
What a Natural Distribution Actually Looks Like
The term "natural anchor text distribution" gets repeated often but rarely defined with precision. In practice, natural means the distribution you would expect if every link were placed by an independent author with no coordination from the site being linked to.
For most established sites, branded anchors form the largest share - often 40 to 60 percent of all inbound links. Naked URLs come next, commonly 20 to 30 percent, because many people simply paste the URL when citing a source. Generic anchors account for another 5 to 15 percent. Partial-match anchors sit in the 5 to 15 percent range, and exact-match anchors - the most powerful individual signal - typically represent fewer than 5 percent of a healthy profile.
An over-optimised profile looks different. Exact-match or partial-match anchors dominate, often at 30 to 50 percent or higher. Branded and naked URL anchors are underrepresented. The distribution tells a story: this site's links were placed by people who were told what anchor text to use.
Google's Penguin Update and the End of Anchor Text Manipulation
Before April 2012, exact-match anchor text manipulation worked reliably. Link builders would acquire dozens or hundreds of links all using the same keyword-rich anchor, and rankings would follow. The technique was mechanical, scalable, and widely understood.
Google's Penguin algorithm update changed the calculus entirely. Penguin analysed anchor text distributions at the domain level and penalised sites whose profiles showed patterns inconsistent with organic link acquisition. Sites that had built aggressive exact-match campaigns saw rankings collapse - sometimes overnight.
Penguin was initially a periodic update, meaning sites could be penalised for months before the next refresh offered relief. From 2016 onward, Penguin became part of Google's core algorithm and began running in real time. The practical consequence is that anchor text manipulation now carries a persistent, continuously evaluated risk rather than a recoverable one-time penalty.
The underlying logic is sound. If a thousand independent authors all linked to the same page using the identical phrase "buy cheap running shoes," the probability that this happened organically is essentially zero. The pattern itself is the signal - not just the individual link.
Best Practices for Internal Link Anchor Text
Internal links are under your full editorial control, which makes them both a significant opportunity and a common source of over-optimisation mistakes. The key principle is to write anchor text for the reader first, then consider the SEO signal it carries.
Descriptive internal anchors work well. If you are linking to an article about how structured data supports modern SEO, an anchor like "how structured data communicates page context to Google" is specific, useful, and carries genuine topical signal. An anchor like "structured data" is acceptable. An anchor like "click here to learn more" wastes the opportunity entirely.
Variety matters even within your own site. If every internal link pointing to a given page uses the same anchor, it looks mechanical - even though you control both sides of the link. Rotate between partial-match variants, branded references, and occasional descriptive phrases that approach the topic from different angles.
Avoid forcing keyword anchors into positions where they read unnaturally. A sentence rewritten to accommodate an exact-match anchor degrades the reading experience and signals editorial manipulation to anyone reviewing the content. The anchor should emerge from the sentence, not the other way around.
Consider context proximity as well. The surrounding text - the words immediately before and after the anchor - contributes to how Google interprets the link. An anchor sitting inside a paragraph about WordPress performance audits carries a different contextual weight than the same anchor dropped into an unrelated sidebar widget.
Best Practices for External Link Building
External anchor text is harder to control because, in legitimate link building, you are persuading an independent author to link to you - not dictating the exact words they use. This constraint is actually an advantage: it forces anchor text diversity that looks natural because it is natural.
When conducting outreach, brief the linking party on the topic of the page you want linked, not the anchor text you want used. If you share a page about Core Web Vitals and user experience, a journalist or blogger who finds it useful will describe it in their own words - producing a partial-match or contextual anchor that carries strong relevance without the manipulation footprint.
For guest posts and sponsored content, the same logic applies. Write the anchor text as you would for your own site - descriptively, naturally, in service of the reader. If an editor changes your anchor to something more generic, accept it. A naked URL or branded anchor from a high-authority domain outperforms a keyword-stuffed anchor from a low-authority one by a considerable margin.
When you do have influence over anchor text - in partner directories, resource pages, or co-marketing content - use branded or partial-match anchors by default. Reserve exact-match anchors for rare, high-authority placements where the topical context is already strong.
Auditing Your Current Anchor Text Profile
Before optimising anchor text, you need an accurate picture of what your current profile looks like. The audit process has two distinct tracks: internal links and external links.
For external links, a crawl tool - Ahrefs, Semrush, or Majestic - will export all inbound links with their associated anchor text. Group anchors by type (exact-match, partial-match, branded, naked URL, generic) and calculate the percentage share of each. Compare that distribution against the natural baseline described earlier. Any category exceeding its expected range warrants investigation.
Pay particular attention to the ratio of exact-match anchors to branded anchors. If exact-match exceeds branded by a significant margin, you are carrying meaningful Penguin risk. The remediation path is to build new links with branded and partial-match anchors - not to disavow existing links unless they are genuinely spammy.
For internal links, a site crawler like Screaming Frog will map every internal link alongside its anchor text. Look for pages that receive multiple internal links all using the same anchor - this is the most common internal linking mistake. Also look for pages receiving only generic anchors ("read more," "learn more") when descriptive alternatives would communicate more clearly to both users and crawlers.
The SEO Analyzer provides a fast, free starting point for auditing on-page and link-related signals without requiring a subscription to a dedicated crawl platform. For a full external link audit, a dedicated backlink tool remains necessary - but for internal link patterns and on-page anchor text hygiene, a page-level audit catches the most common issues quickly.
One useful benchmark: run the same audit on a competitor that ranks well for your target terms. Their anchor text distribution reveals what Google has already accepted as natural in your niche. Match the pattern, not the exact anchors.
Good Anchors Versus Bad Anchors: Concrete Examples
Abstract principles become clear with specific comparisons. Consider a page targeting the topic of WordPress performance optimisation:
| Anchor Text | Type | Assessment |
|---|---|---|
| WordPress performance optimisation | Exact-match | High signal, high risk if overused - reserve for rare high-authority placements |
| tips for speeding up your WordPress site | Partial-match | Strong topical signal, reads naturally, low manipulation footprint |
| Signocore's performance guide | Branded partial-match | Builds brand authority alongside topical relevance - an underused combination |
| https://signocore.com/articles/core-web-vitals | Naked URL | Neutral signal, adds diversity, common in citations |
| click here | Generic | No topical signal, poor UX, wasted opportunity |
| this article | Generic | Marginally better than "click here" - still contributes nothing to relevance |
The table makes the tradeoff visible. Exact-match anchors are not wrong - they are simply high-risk when overused. Partial-match and branded anchors deliver most of the topical benefit with a fraction of the footprint. Generic anchors are the option of last resort, not a safe default.
Anchor Text as a Signal, Not a Lever
The shift in how SEO professionals should think about anchor text is a shift from mechanical control to editorial quality. The sites that rank well over the long term are not the ones that engineered the perfect keyword ratio in 2015 - they are the ones that created content worth linking to, earning anchors that describe the content accurately because the authors chose those words themselves.
That does not mean anchor text optimisation is passive. Internal links are fully within your control and should be treated with the same editorial care as any other content decision. External anchor text can be influenced through the quality and clarity of the content you produce and the context you provide to those linking to it. The discipline is knowing the difference between influence and manipulation - and staying firmly on the right side of that line.
