The shape of a natural anchor text profile is well-documented. The shape of a manipulated one is documented too. The gap between them is what Penguin watches for.
Google’s Penguin algorithm launched in April 2012 specifically to address anchor text manipulation. Before Penguin, the path to ranking for a commercial keyword was straightforward: build a lot of links with that keyword as the anchor text, and rankings would follow. The tactic worked well enough that it dominated link building strategy for years.
Penguin changed the math. The algorithm identifies anchor text patterns that don’t occur naturally and suppresses rankings for sites showing those patterns. In September 2016, Google folded Penguin into the core algorithm and made it operate in real time. The algorithm doesn’t get “refreshed” any more; it’s always evaluating, every time Google re-crawls a site’s backlink profile.
The result for 2026: anchor text optimization is no longer about getting your target keywords into anchors. It’s about maintaining a distribution that looks like what would naturally emerge if real people were genuinely linking to your content, while still capturing enough relevance signal to support ranking. The balance is narrow and the cost of getting it wrong is substantial.
What follows is the breakdown of what natural anchor profiles look like. It covers what manipulation patterns trigger Penguin, and how to think about anchor strategy when you have partial control over what other sites use as anchor text.
The six categories of anchor text:
Every backlink anchor falls into one of six categories. The mix across all the anchors pointing at a page is the page’s anchor profile.
Branded anchors use the brand name or domain. “Apple,” “Nike,” “yoursite.com.” These are the most common anchors in genuinely organic link profiles. People writing about a brand naturally reference it by name. Branded anchors don’t pass strong keyword relevance for non-branded queries but signal entity recognition.
Naked URL anchors are the bare URL displayed as the anchor text. “https://yoursite.com/page-name.” These appear when authors copy and paste links without adding descriptive text. The pattern is genuinely common in casual references, social posts, and copy-paste citations.
Generic anchors are non-descriptive phrases like “click here,” “read more,” “this article,” “visit this page,” or “learn more.” They provide little keyword relevance. They reflect common natural linking behavior, especially in resource pages and casual references.
Partial-match anchors include variations of target keywords without using the exact phrase. If a page targets “project management software,” partial-match anchors might include “project management tools,” “managing projects,” or “software for project teams.” They provide keyword relevance with natural variation.
Exact-match anchors use the precise target keyword. If a page targets “project management software,” the exact-match anchor is “project management software.” These pass the strongest keyword relevance signal. They also trigger the most algorithmic scrutiny when overused.
Image anchors apply when a hyperlink is wrapped around an image. The alt text of the image functions as the anchor text. The category is small for most sites but matters for image-heavy pages.
A natural anchor profile distributes across these categories in proportions that vary by industry, content type, and the brand’s organic mention patterns. The proportions for B2B SaaS look different from the proportions for e-commerce, which look different from the proportions for local services. But within each category, the general patterns of natural distribution have been documented enough that significant deviations stand out.
The natural distribution benchmarks:
Industry analyses and SEO research have converged on rough benchmarks for natural anchor text distribution. The benchmarks vary across sources but the consensus ranges:
Branded anchors: 30-50% of total. Higher for established brands with strong name recognition; lower for newer brands with less organic mention.
Naked URL anchors: 10-25%. Higher for B2B and technical content where citations are common; lower for consumer-facing content.
Generic anchors: 10-20%. Stable across most industries.
Partial-match anchors: 10-25%. Higher for content that addresses specific topics in detail; lower for general brand pages.
Exact-match anchors: 1-5%. The lowest natural percentage. Genuinely organic content rarely produces high concentrations of exact-match anchors because real people write in their own words.
Topically related anchors: 5-10%. Related terms that aren’t exact or partial matches but appear in natural discussion of the topic. (Some SEO writing still calls these “LSI anchors,” but the term is a misattribution; Google’s John Mueller has confirmed repeatedly that LSI as an information retrieval technique isn’t part of Google’s ranking systems. The category is real even though the label is technically wrong.)
Image and other: 5-15%. Filler category covering image links, mixed anchors, and edge cases.
The benchmarks describe what natural profiles tend to look like, not what brands should target. The difference matters. A profile that hits the benchmarks artificially (by deliberately building anchors to match the proportions) still has the manipulation signature of any other engineered link profile. The benchmarks are descriptive, not prescriptive.
What actually produces natural distributions: real editorial linking from sites that genuinely reference the content for reader benefit. The distributions emerge as a byproduct. They can’t be reverse-engineered through deliberate construction without creating the patterns Penguin detects.
What over-optimization looks like:
The patterns Penguin identifies and suppresses follow recognizable shapes. The most consistent triggers:
Exact-match concentration above 8-10% of total anchors. The threshold varies by industry and competitive context, but commercial keyword anchors above 10% of a page’s profile is reliably risky. Above 15-20% is in penalty territory for most cases.
Repeated identical anchors across many domains. A page receiving 30 backlinks where 25 of them use the exact same anchor text indicates coordinated placement rather than organic linking. Natural linking produces variation; coordinated placement doesn’t.
Anchor text concentration disproportionate to the page topic. A page about “wedding photographer Boston” with 60% of its anchors being “wedding photographer” looks built for the keyword rather than earned through editorial coverage of weddings.
Commercial keyword anchors with high concentration. Anchors like “buy cheap SEO services” or “best affordable web hosting” are unambiguously commercial. High concentrations of commercial-intent anchors indicate paid placement or link scheme participation.
Sudden anchor distribution shifts. A profile that’s been 60% branded for years and suddenly shifts to 30% exact-match within a few months signals an active link-building campaign.
Anchor velocity bursts. A page acquiring 50 new backlinks with similar anchor text in a single week looks unnatural regardless of how the anchors are distributed across the new links.
Geographic concentration patterns. A page for a local business in Chicago receiving anchors that all reference “Chicago [industry]” in slight variations suggests scheme participation, even when individual anchors aren’t identical.
Anchors that don’t match the destination content. If the destination page is about residential plumbing but the incoming anchors are about commercial roofing, the mismatch indicates either error or manipulation.
Penguin’s evaluation isn’t deterministic. The algorithm considers patterns in context. A 12% exact-match ratio might be fine on a page where the topic is genuinely defined by that exact phrase. The same ratio could be problematic on a page where the topic has many natural phrasings.
The internal vs. external anchor distinction:
A consistent area of confusion is the difference between internal anchor text rules and external anchor text rules. They operate differently because the control situation is different.
Internal links are under the site’s full control. The site owner chooses the anchor text for every internal link. Google understands this and treats internal anchors differently from external anchors.
For internal links:
Descriptive, keyword-relevant anchor text is appropriate and beneficial. Google has confirmed there’s no penalty for internal anchor text optimization. The site is communicating its own architecture; clear, descriptive anchors help both users and Google understand the site’s structure.
Avoid “click here” and “read more” for important internal links. These generic anchors waste the relevance signal that internal linking can provide.
Vary the anchor text across multiple internal links to the same destination. Using exactly the same anchor on every internal link to a page is slightly less effective than varying the anchor naturally. The variation helps Google understand the page’s topic from multiple angles.
For external links (backlinks pointing to the site):
The site has limited or no control. Webmasters of linking sites choose their own anchor text. The result is naturally varied.
Distribution should look organic. The benchmarks above apply. Concentration of exact-match anchors on commercial pages is the primary risk.
Editorial links produce natural distributions. Manipulated links require careful engineering to avoid triggering Penguin, which is the whole reason the algorithm catches them.
The implication: the rules for “how to write anchor text” depend on which direction the link is going. Internal: descriptive, varied, keyword-aware. External: whatever the linking site chose, with the overall profile distribution monitored at the page level.
What you can do when you have partial control:
Some link acquisition situations give the brand partial influence over the anchor text. The most common:
Guest post contributions where the contributing author has input into the anchor text of links in their bio or within the article.
Sponsored placements where the brand is paying and can specify anchor text (subject to required nofollow/sponsored attributes).
Directory listings where the brand fills in the form fields.
Editorial relationships where the publisher will accommodate the brand’s preference if asked.
What to do with the partial control:
- Skew toward branded and partial-match anchors. The exact match anchor temptation is the trap; resisting it preserves the overall profile.
- Suggest natural-sounding options when asked. “Project management software” as a partial-match works better than “the best project management software for agencies.”
- Vary the anchor across different placements. If contributing five guest posts that link back to the same page, use five different anchor variations rather than the same one five times.
- Match the anchor to the surrounding content. The anchor should fit the sentence it’s in. An anchor that requires the surrounding text to be twisted around it stands out.
- Avoid commercial keyword anchors entirely on guest contributions. “Best [product] for [use case]” reads as paid placement regardless of whether payment was involved.
- Accept the editor’s choice when they choose differently. If the publication’s editor wants to change the anchor, the change is usually correcting an over-optimization the brand had proposed.
The pattern: even with influence over anchor text, the right strategy mostly involves not pushing for the optimization that the algorithm flags. The links that survive Penguin scrutiny are the ones where the anchor wasn’t optimized aggressively in the first place.
The page-level vs. site-level question:
Anchor text ratios apply at the page level, not the domain level. A site can have an overall healthy distribution while specific pages have problematic patterns.
The mechanism: Penguin evaluates anchor patterns per landing page. The homepage might have 85% branded anchors and the services page might have 25% exact-match commercial anchors. The homepage looks fine; the services page is at risk.
The implication for analysis:
Anchor audits should examine each significant landing page separately. Sitewide averages can hide page-level problems.
The pages most at risk are typically commercial pages targeting specific keywords. Service pages, product pages, and conversion-focused landing pages tend to attract the link-building attention that produces over-optimization.
Lower-risk pages (the homepage, blog posts, resource pages) often look healthy even when commercial pages are in trouble.
The audit workflow:
Export all backlinks for the site through Ahrefs, Semrush, or similar tools.
Group the backlinks by destination URL. The grouping reveals which pages have the most external links.
For each page with significant backlinks, calculate the anchor distribution across categories.
Flag any page where exact-match anchors exceed 8-10% of total.
Flag any page where a single anchor variant repeats more than 5 times across different domains.
Flag any page where the distribution looks substantially different from the benchmarks.
The flagged pages are the priority for remediation, regardless of how the site’s overall profile looks.
Recovery from anchor over-optimization:
When a page has anchor text issues that correlate with ranking suppression, the recovery process has documented steps. The process takes months and requires discipline.
Phase 1: Stop the bleeding. Halt all link building that adds exact-match or commercial keyword anchors to the affected pages. The continuation of the pattern prevents recovery regardless of what other steps are taken.
Phase 2: Build offset links. New links to the affected pages should use branded, generic, or naked URL anchors exclusively. The offset links dilute the over-optimized ratio without requiring the existing manipulative links to be removed or disavowed.
Phase 3: Update internal linking. Internal links pointing to the affected pages can be adjusted to use more varied anchor text. The change provides additional dilution and is fully under the site’s control.
Phase 4: Wait for recrawl and reevaluation. Penguin operates in real time, but Google’s recrawl frequency varies. The recalibration of the page’s anchor profile takes weeks to months as the new offset links are crawled and incorporated into the evaluation.
Phase 5: Selective disavow if needed. For pages where the over-optimization came from clearly manipulative sources (paid placements, low-quality networks, obvious scheme participation), targeted disavow of the worst offenders can help. The disavow file should be conservative; broad disavow often produces additional damage.
The realistic timeline: 3-6 months for the recovery to show measurable ranking improvement. Longer for severe over-optimization or for pages that lost rankings deeply.
The frequent mistake: trying to accelerate recovery by removing or disavowing many existing links aggressively. The acceleration usually doesn’t work and frequently makes the situation worse by removing some links that were contributing positive signal. The patient approach of dilution through new clean links outperforms aggressive cleanup in most cases.
The AI search overlay:
Anchor text strategy in 2026 has gained an additional consideration through AI search engines.
AI engines like ChatGPT, Perplexity, and Google’s AI Overviews don’t weight anchor text the same way Google’s core ranking does. They use semantic retrieval, which means the words around the link matter more than the link’s anchor text. A descriptive sentence that contains the link, even with a generic anchor, can be more effective for AI citation than an exact-match anchor in a thin context.
The split strategy that emerges:
For Google safety, keep exact-match anchors below 5% of the external profile. The restraint protects against Penguin.
For AI citation, focus on the context surrounding the link rather than the anchor itself. Sentences that clearly describe what the destination contains, even with branded or generic anchors, get retrieved by AI systems more effectively than keyword-stuffed anchors with thin context.
The two approaches reinforce each other. The anchor strategy that protects against Google over-optimization (branded, generic, naked URL) also produces the context-rich placements that AI systems prefer. The conflict that existed in earlier years between “optimized for Google” and “natural for readers” has largely resolved into a single approach that works for both.
The strategic position:
Anchor text in 2026 is a practice of restraint. The optimization that worked in 2010-2014 (build many links with target keyword anchors) doesn’t work anymore. The optimization that works now is the absence of obvious optimization. The goal: a profile that looks like what genuine editorial linking would produce, with enough partial-match variation to support relevance for the target queries.
The brands that consistently maintain healthy anchor profiles share characteristics:
They earn most of their links through editorial coverage rather than building them deliberately. Editorial links produce natural distributions automatically.
Brand mentions in their own content and PR materials get used more than commercial keywords. The mentions get picked up and reproduced in coverage, which feeds the branded anchor share organically.
Anchor profiles get audited quarterly. Catching over-optimization early allows correction before ranking damage accumulates.
Exact-match anchors never get pushed when partial control over guest posts or sponsored placements is available. The discipline of not optimizing aggressively is what keeps the profile safe.
Content quality earns the links rather than optimization tactics maximizing keyword relevance. The content-driven approach produces sustainable results; the optimization-driven approach produces short-term gains followed by Penguin corrections.
The brands that struggle with anchor profiles share opposite characteristics. They built links aggressively in earlier eras and never corrected the resulting profiles. Third-party services that promise specific keyword anchors stay in rotation. Ranking for specific terms gets prioritized over building broader topical authority. Each link acquisition becomes an opportunity to maximize keyword relevance rather than a piece of a broader profile.
The anchor profile is one of the SEO indicators that most reliably reveals how a site’s link building has been conducted. Healthy profiles indicate sustainable practice. Over-optimized profiles indicate either historical mistakes that need correction or ongoing practices that need to stop. The difference shows in rankings consistently and predictably enough that anchor analysis is one of the first diagnostic steps in any backlink audit worth running.