What should your next ten links say?

The number nobody wants to show you

Suppose a page has forty links and twelve of them use the exact target keyword. That is thirty per cent exact match, against a balanced target of seven. The obvious move is to build more links with safer anchors and dilute it.

The arithmetic says otherwise. To get twelve exact-match links down to seven per cent of the profile, the profile has to reach 172 links. That is 132 more — not a campaign, a programme running well over a year. A ten-link buy moves the ratio from thirty per cent to twenty-four, which is still three times the target and still reads as built.

That calculation takes a minute on paper and almost nobody does it, which is why so many campaigns are sold as fixes for a problem they cannot arithmetically fix. When your inputs produce that answer, this tool gives it to you plainly, and points at removal instead — because we would rather tell you not to buy links than sell you links that will not work.

The seven types, and what each is for

Anchor classification is usually presented as a taxonomy to memorise. It is more useful as a set of jobs. Branded anchors establish that a real business is being referred to. Partial matches carry most of the topical signal at a fraction of the risk. Naked URLs and generics look like nothing, and a profile with none of them looks assembled — the absence of noise is itself a signal.

The bucket most built profiles are missing entirely is topical: natural phrases about the subject that do not contain the keyword at all. Real editors write those constantly. Link builders almost never buy them, because they feel like a wasted placement. That gap is one of the more legible fingerprints of a paid profile.

Why the targets are a choice and not a fact

We would rather hand you a single safe number, but it does not exist. Guidance published in the last two years puts the exact-match ceiling at five per cent, at eight, and at twenty, while measurement of pages that actually rank puts the observed figure near thirteen. Those cannot all be right, and the disagreement is not going to resolve because Google does not publish a threshold.

So the tool asks what you are optimising for. A site recovering from a manual action should sit well inside the strictest reading. A site with no history in a low-scrutiny niche can reasonably sit near what the ranking pages do. Both are defensible; pretending there is one answer is not.