Anchor text distribution that is heavily exact-match is a pattern link schemes produce, and Google’s link spam systems — the descendants of Penguin, folded into core ranking since 2016 — largely respond by ignoring the links rather than penalising the site. Natural profiles are brand-heavy with diverse anchor variety. The fix is auditing distribution, identifying excessive exact-match patterns, and shifting new-link efforts toward brand and varied anchors. Connect to backlink strategy.
Healthy profile for most sites:
Branded anchors (Acme, Acme Corp): 30-50%
Naked URLs (example.com): 10-20%
Generic ("click here", "read more"): 5-15%
Partial-match ("Acme CRM tool"): 10-20%
Exact-match ("CRM software"): 5-15%
Page title / heading text: 5-10%
Other variations: 5-10%
Rule of thumb, not a threshold: exact-match > 20-25%
Google publishes no anchor ratio and no cut-off — every percentage on this page is community convention inferred from observation, not a documented limit. Treat a high exact-match share as a symptom worth investigating, not as a trigger that has been pulled.
Ahrefs / Semrush / Majestic show anchor text reports. Group anchors into categories above. Calculate percentages. Common problems found: - Single exact-match phrase > 30% (no "trigger" exists, but this is a pattern only link-buying produces) - Naked URLs < 5% (looks artificially curated) - Generic anchors < 2% (no natural variation) - Brand < 20% (looks like you're not real brand)
If exact-match is 40% and should be 15%: Option 1: Add diverse new anchors (slow) - Build new links with brand, naked URL, partial-match - Wait 6-12 months for ratio to drift - Safer but slower Option 2: Remove some exact-match links (faster) - Outreach to remove excessive exact-match links - Or disavow if removal fails - Risk: losing actual authority along with anchor Option 3: Vary anchors on existing links (negotiation) - Request publishers update anchor text - From "CRM software" to "Acme CRM software" or naked URL - Most editors will accommodate reasonable requests
Per outreach campaign, plan anchor distribution:
10 outreach campaigns → 10 placements:
4 brand anchors (Acme)
2 naked URLs (example.com)
2 partial-match (Acme CRM tool)
1 exact-match (CRM software)
1 page title / generic
This gradually shifts overall ratio toward natural.
Your own internal linking contributes to anchor signal. If every internal link to /crm-software/ uses "CRM software" anchor, that contributes to over-optimisation. Vary internal anchors too:
For /crm-software/, internal anchors might be: "our CRM" "Acme's CRM platform" "the CRM software guide" "see CRM features" "CRM software" Same destination, varied anchors. Looks natural, distributes authority across keyword variants.
Quarterly: - Re-audit anchor distribution - Track exact-match percentage trend - Investigate spikes (negative SEO indicator?) - Adjust outreach anchor mix - Update internal linking on key pages
Every percentage on this page — including the ones in the table above — is community convention. Google has never published an anchor text ratio, a threshold, or a cut-off, and no leaked document has ever contained one. These figures come from people looking at sites that did well and sites that did not, and inferring.
That does not make them useless. It changes what they are: a description of what a natural profile tends to look like, rather than a target you are being scored against. And the distinction has real consequences for what you do next.
If you believe there is a line at 25%, you will try to engineer your way back under it — disavowing links, chasing publishers for anchor edits, building filler links with brand anchors to dilute the ratio. That is a lot of work aimed at a number nobody is measuring.
If you understand the ratio as a symptom, you ask the more useful question: why is my exact-match share 40%? There is essentially only one answer, and it is that somebody was choosing the anchors — which means links were bought, exchanged, or placed. The ratio is the fingerprint. Fixing the fingerprint while keeping the habit changes nothing.
The word penalty does a lot of unearned work in link advice, and it is worth separating two things that are constantly conflated.
Algorithmic devaluation (common) - Google's link spam systems distrust the links - The links stop counting - Nothing is "applied" to your site - Outcome: you are back where you started, minus the money Manual action (rare) - A human reviewer at Google acts on your site - It is reported in Search Console - It requires a reconsideration request - If you have not received one, you do not have one
Penguin has been part of core ranking since 2016, and its descendants largely operate by ignoring links they do not trust rather than by punishing the sites that received them. That is a mercy and a warning at once: the downside of a bad link is usually that it does nothing, and the money spent on it is simply gone.
Which reframes the disavow question in option 2 above. Disavowing is not a way to lift a punishment that, in most cases, was never applied. It is a way to tell Google to ignore links it very likely already ignores — while risking the loss of the ones that were genuinely helping. Reserve it for links you know were paid for or placed as part of a scheme, and for the rare case of a manual action naming them.
Every option in section 3 is a way of adjusting a ratio after the fact. None of them addresses the thing that produced it, and a profile corrected without a change in habit drifts straight back.
A natural anchor profile is not something you build. It is a byproduct of links you did not choose the anchors for — which is another way of saying links somebody gave you because your page was worth citing. People linking naturally use your brand name, the page title, a bare URL, or a scrap of the sentence they were writing. That is exactly why the natural pattern looks messy: nobody was optimising it.
So the fix that lasts is upstream of anchor text entirely: publish something specific enough that other people want to reference it, and let them describe it however they like. The distribution takes care of itself, and you never have to audit it again.
Almost all anchor advice is about links other people give you, which is the half you can influence least. Your internal links are the half you own completely, and they are routinely wasted.
Two errors dominate. The first is anchor text that says nothing — "click here", "read more", "learn more" — which tells a reader nothing about where they are going and tells a parser nothing about what the destination covers. The second is the opposite: the same exact-match phrase used as the anchor for every internal link to a money page, on the theory that repetition strengthens the signal. It does not, and a site whose internal anchors are uniformly optimised is displaying the same curation tell that gives away a bought external profile.
The rule that resolves both: describe the destination honestly, in the words the sentence naturally wants. Sometimes that is the exact keyword. Sometimes it is the page title. Sometimes it is a phrase halfway between. Written that way, internal anchors vary on their own — and they do the one job anchors are genuinely good for, which is telling a reader what they will get if they click.
This also matters for accessibility, which is the argument nobody makes and probably should. A screen-reader user can pull up a list of every link on the page and navigate by it. A list of fourteen entries reading "click here" is useless. Link text is an interface element before it is an SEO signal, and writing it well serves both.
Section 6 tells you to investigate spikes as a possible negative-SEO indicator, and it is worth being calm about what that means, because worry here consumes more time than the threat justifies.
Somebody pointing a few thousand spam links with identical exact-match anchors at your site is a real thing that happens. What it usually achieves is nothing. Google's link spam systems are built on the assumption that a site does not control who links to it, so the ordinary response to an obviously artificial burst of links is to ignore them — which is the same response those links would get if you had bought them yourself.
So the reasonable posture is to look, note it, and carry on. Check whether your rankings actually moved, because a spike in a backlink tool with no movement in the results is a spike in a backlink tool. If nothing changed, nothing changed.
Distribution matters less below ~50 backlinks. Focus on getting more links, varied naturally. Distribution audit becomes important above 100 backlinks.
Less but still partial signal. Google ignores nofollow for PageRank but observes anchor patterns for context. Still avoid heavy exact-match on nofollow links.
Risky — you'd lose authority too. Better: shift the ratio by adding diverse new links. Only disavow if specific exact-match links are clearly spammy.
Rarely. Brand-heavy is natural — most organic links use brand or naked URLs. Concern starts above 70-80% brand if it looks unnaturally curated.
No. Google has never published an anchor text ratio, a threshold, or a cut-off, and every percentage in guidance like this — including on this page — is community convention inferred from observing sites, not a documented limit. Treat the figures as a way of describing what a natural profile tends to look like, and never as a number you are being scored against. Sites are not penalised for crossing a line, because there is no line; they get into trouble for buying links, and a distorted anchor profile is one of the marks that habit leaves.
Usually not. Google's link spam systems — the descendants of Penguin, folded into core ranking since 2016 — largely work by ignoring links they distrust rather than by penalising the site that received them. So the common outcome of an unnatural profile is not a punishment applied to your domain; it is that the links you bought simply stop counting, and you are back where you started minus the money. Manual actions do exist, and they are rare and reported in Search Console. If you have not received one, you almost certainly do not have one.
Nofollow became a hint rather than a directive in 2019, which means Google may or may not use those links for ranking and crawling at its discretion — so the honest answer is that nobody outside Google knows how they are weighted. What follows practically is unchanged: a pile of nofollow exact-match anchors from paid placements is still evidence of a link-buying habit, and the right response to that evidence is never to build the habit rather than to audit the tags.
aiwebpageseo.com is a data-driven SEO and AEO (Answer Engine Optimisation) platform providing a free suite of technical website tools. Rather than relying on AI-theorised assumptions, the platform analyses live URL performance, delivering objective diagnostics, page speed metrics, CLS debugging, and site crawl data alongside actionable technical tutorials.