What Is a Healthy Anchor Text Ratio?

Google publishes no anchor text ratio, so any figure you see is a working benchmark rather than a rule. In practice a healthy profile is roughly half branded and naked-URL anchors, a quarter generic or partial-match phrases, and well under 10% exact-match commercial keywords. What matters is that the distribution looks accidental — real editorial links rarely repeat the same money phrase.

That last sentence is the whole idea, and it's worth sitting with before you go chasing percentages. Nobody at Google decided that 8% exact-match is safe and 11% is not. The ratios circulating in SEO forums are reverse-engineered from what naturally-linked sites happen to look like. They are useful as a mirror — does my profile resemble one that grew on its own? — and useless as a target to hit.

The working benchmark

If you want a number to sanity-check against, this is the shape most healthy profiles fall into:

Anchor type	Typical share	Example
Branded	35–50%	"Backlinkster", "Backlinkster's guide"
Naked URL	10–20%	backlinkster.com
Generic	10–20%	"this guide", "here", "read more"
Page title / long phrase	10–20%	"What Is a Healthy Anchor Text Ratio?"
Partial match	5–15%	"anchor text distribution explained"
Exact match	1–5%	"healthy anchor text ratio"

Two things surprise people about this table. The first is how small the exact-match slice is — most site owners building links deliberately land at three or four times that. The second is how large the "useless" categories are. Branded, naked-URL and generic anchors carry almost no keyword signal, and together they're the majority of a natural profile. That's not a flaw in the profile; it's the fingerprint of links a human editor chose to place.

What is anchor text covers the categories themselves if any of those terms are unfamiliar.

Why there is no official number

Google's link spam policy describes the behaviour it acts on — links intended to manipulate ranking — not a threshold it enforces. There is no published cutoff, and there almost certainly isn't a fixed internal one either, because the right distribution genuinely varies:

A brand-name product with high search demand for its own name will run 60%+ branded anchors without anything being wrong.
A generic-name site ("City Plumbing Supplies") will show far more exact-match by pure accident, because its brand name is the keyword.
A new site with 12 links has no meaningful distribution at all. Three exact-match anchors out of 12 is 25% and means nothing.

This is why ratio-chasing goes wrong. Someone reads that exact-match should be under 5%, sees they're at 14%, and starts disavowing links — often links that were helping them. How to disavow bad backlinks covers why that's usually the wrong move.

What actually triggers a problem

The signal isn't a percentage. It's a pattern that couldn't have happened by accident. Three combinations do the real damage:

One phrase repeated across unrelated domains. Forty sites, in four different industries, all linking with "best crm software" in the same sentence structure. No editor writes that way. This is the classic Penguin-era footprint and it's still the clearest tell.

Exact-match outnumbering your brand. If more links call you "cheap car insurance quotes" than call you by your own name, the profile is inverted relative to every organically-linked site on the web.

Commercial anchors on site-wide placements. A keyword-rich anchor in a footer or sidebar template, replicated across thousands of pages, is both low-value and conspicuous. See are footer and sidebar links good for SEO.

Note what's not on that list: a handful of exact-match anchors from relevant sites, foreign-language anchors, scraper sites copying your own internal anchors, or a single guest post where you picked the anchor. Those are ordinary.

How to measure your own distribution

Fifteen minutes, no paid tool required:

Export your anchors. Search Console → Links → Top linking text gives you Google's own view, which is the view that counts. A third-party crawler adds coverage.
Deduplicate by domain first. One site linking 900 times with the same anchor is one data point, not 900. Skipping this step is the single most common reason people think their ratio is broken. Referring domains vs backlinks explains the distinction.
Bucket each anchor into the six categories in the table above. A spreadsheet and twenty minutes beats any tool's automatic classification, which routinely mislabels partial matches.
Calculate shares and compare to the benchmark — as a mirror, not a scorecard.
Separate what you built from what arrived on its own. This is the step that turns the exercise into a decision. You control the first group and can only observe the second.

If your total is under about 50 referring domains, stop here. The numbers aren't stable enough to act on.

Fixing a skewed profile

You almost never fix a ratio by removing links. You fix it by changing what you build next and letting the denominator grow.

Stop building the anchor that's over-represented. Immediately and completely.
Dilute deliberately. Make the next 20–30 links branded, naked-URL, or page-title anchors. A profile at 18% exact-match with 60 domains drops under 10% once you add 50 branded links — without touching anything that already exists.
Vary the phrasing, not just the category. Ten links reading "anchor text guide", "guide to anchor text" and "this anchor text guide" still look coordinated. Real editors write differently from each other because they're different people.
Remove at source only what you placed and paid for. Anything you can email a webmaster about and have quietly changed is worth changing. Anything you can't is a candidate for dilution, not disavowal.

Time is on your side here. Anchor distribution is a rolling average over your whole profile, so consistent boring link building moves it without any cleanup work at all.

Where exchanges fit

Link exchanges are one of the few places you genuinely choose your own anchor, which makes them either the fastest way to over-optimize or the easiest way to correct a profile — entirely depending on discipline.

The discipline is simple: don't spend your one controllable anchor on your money keyword every time. Backlinkster places every swap as an in-content link inside a real sentence, which naturally pushes anchors toward page titles and descriptive phrases rather than bare keywords, and it verifies each placement stays live so your distribution reflects links that actually exist. Free accounts get five swaps a month; see the plans.

Frequently asked questions

What is a healthy anchor text ratio? Roughly 35–50% branded, 10–20% naked URL, 10–20% generic, 10–20% page-title or long-phrase, 5–15% partial match, and 1–5% exact match. These are observed patterns from naturally-linked sites, not thresholds Google publishes or enforces.

What percentage of anchor text should be exact match? Most healthy profiles sit between 1% and 5%. Under 10% is comfortable. The risk isn't the percentage itself but the pattern — the same commercial phrase repeated across unrelated domains is what looks manufactured, whatever the arithmetic says.

Does Google penalize over-optimized anchor text? Google's link spam systems devalue links intended to manipulate ranking, and a heavily exact-match profile is one of the clearest indicators of that intent. The usual outcome is quiet devaluation rather than a manual action, so you lose the value of the links without any notification.

How do I check my anchor text ratio for free? Search Console's "Top linking text" report under Links is free and shows Google's own view. Export it, deduplicate by referring domain, bucket the anchors by type, and calculate the shares in a spreadsheet.

Should I disavow links because of anchor text? Almost never. Anchor distribution is fixed by changing what you build going forward, not by removing what exists. Reserve the disavow file for a manual action or for paid links you can't remove at source.

Does anchor text ratio matter for internal links? Less, and differently. Google understands you control your own site, so internal anchors are read as navigation and description rather than as third-party endorsement. Descriptive internal anchors help; there's no ratio to police.

The bottom line

Treat the benchmark as a mirror. Export your anchors, deduplicate by domain, bucket them, and ask one question: could this distribution have happened without me? If yes, you're fine at almost any percentage. If no, stop building that anchor and dilute forward — that fixes far more profiles than any disavow file ever has.