AnchorTextRatiosAreOverthought

There is a genre of SEO advice that specifies anchor text distribution to the percentage point. Twenty percent branded, fifteen percent exact match, the rest naked URLs and generics. It is stated with total confidence and I have never seen it survive contact with real data.

Across the recovery and clean-up work I have run since 2023, I have anchor distributions for a few dozen profiles alongside what happened to them. It is not a study and I would not present it as one — but the pattern in it is not subtle.

Profiles with heavy exact-match ratios that recovered fine. Profiles with textbook-clean ratios that got hit anyway. The ratio on its own predicted very little.

Every profile that took damage shared something the ratio was only ever a downstream symptom of.

• Anchors that clustered in time. Forty links carrying the same phrase inside six weeks, then nothing. The distribution is unremarkable. The timing is a confession.
• Anchors that clustered by source type. The same phrase arriving repeatedly from the same category of low-effort site.
• Anchors nobody would write. Nobody links out with the words "best crm software for small business 2024". A human writing a sentence does not produce that string.

The ratio is the shadow. The acquisition pattern is the object casting it.

I do not manage a ratio, and I do not let clients set one as a target. Anchor text is written by whoever writes the sentence, and the sentence is about the point being made rather than the phrase we would like to rank for.

That produces a distribution that is messy, uneven, and different on every engagement — which is the same thing that happens to sites that never thought about it at all. That resemblance is the entire objective.