Anchor text vs. co-occurrence text: optimizing the wrong field

Anchor text vs. co-occurrence text: optimizing the wrong field

Practitioners pour effort into the anchor string itself while ignoring the surrounding text — the words around the link. Evidence from patents and analysis suggests the surrounding context may carry relevance signal independent of the anchor, which reframes the optimization target.

Google has long-held patents describing the use of text near a link, not just the anchor, to determine topical relevance. If the paragraph around a branded anchor is rich with topical terms, the link can transmit relevance without the anchor itself being exact-match.

The tactic: instead of forcing an exact-match anchor (high risk), use a branded or natural anchor inside a paragraph that naturally contains your target terms. You get relevance from co-occurrence while keeping the anchor itself low-risk.

On one hand, this decouples relevance from anchor over-optimization elegantly — the riskiest field stays clean. On the other, the weight Google places on surrounding text versus the anchor proper is undisclosed and likely modest relative to the anchor.

Limitation: co-occurrence weighting is inferred from patents that may not reflect current production systems, and from correlation studies that cannot isolate it.

My practice: write the link's host paragraph deliberately, treating it as a relevance surface, not just the anchor.

Open question: what is the relative weight of surrounding text vs. the anchor itself in current ranking systems?