Two domains, mirrored link sets, divergent anchors

Two domains, mirrored link sets, divergent anchors

A rare quasi-experiment surfaced when an agency built links for two competing clients in the same niche from overlapping publisher lists in 2023 — roughly 65% source overlap, comparable domain authority, launched within the same quarter.

Client A received a profile engineered at ~12% exact-match, 50% branded, rest generic/URL. Client B's profile drifted to ~34% exact-match because the vendor reused commercial anchors more aggressively.

At 10 months, Client A held a median position of 5 across the shared keyword set; Client B sat at 12, despite slightly more total links.

Because the source pool overlapped heavily, this approximates holding link quality constant while varying anchor distribution — methodologically stronger than most public cases.

On one hand, this is close to the controlled comparison the field lacks. On the other, two clients are not identical: different on-page, different brand search volume, different internal linking. Those could carry the gap entirely.

Limitation: n=2, unblinded, agency-reported, every incentive to credit the anchor strategy.

Confidence: suggestive, not conclusive — exactly the kind of case that should motivate a real test, not end the argument.

Open question: how much of A's lead traces to brand-search demand rather than the anchor ratio?