[ Case study ]
A manual-action penalty arrived for a link scheme the previous vendor had run — thousands of directory and blog-network links, mostly in a foreign language, pointing at pages the agency had never written. Traffic fell off a cliff and new business followed it.
CLIENT a digital marketing agency — FOCUS Separate the salvageable from the toxic
Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.
Marketing agencies live by their own search presence, so when a manual-action penalty hit this one — for a link scheme a previous vendor had run — the damage was both revenue and reputation. Thousands of directory and blog-network links, mostly in a language the agency doesn't speak, pointed at pages the agency never wrote. Traffic fell off a cliff and new business followed it down. The team kept serving clients through the recovery; the managing director made time in two fixed slots a week, and that was all the calendar could give.
A manual-action penalty arrived for a link scheme the previous vendor had run — thousands of directory and blog-network links, mostly in a foreign language, pointing at pages the agency had never written. Traffic fell off a cliff and new business followed it.
We proposed separating the salvageable from the toxic before touching anything: the backlink profile classified into genuinely earned links, payable-looking directory noise, and pure network spam, with the earned set documented as the recovery target. The disavow file would then cover network patterns and junk directories precisely — legitimate-but-weak links kept and improved rather than burned. The reconsideration request would present that classification as evidence of judgment. In parallel, two original research pieces from the agency's own campaign data would earn real links while the review ran.
Just as important is what we ruled out, and why:
The backlink profile was classified — genuinely earned links, payable-looking directory noise, and pure network spam — with the earned set documented as the recovery target.
The disavow file covered the network patterns and junk directories precisely; legitimate-but-weak links were kept and improved rather than burned.
Two original research pieces from the agency's own campaign data earned coverage while the reconsideration request was prepared.
Delivered by the growth pod — authority specialist over 9 weeks, with working increments reviewed with the client every week.
Obstacle
The earned-link baseline had never been documented, and reconstructing 'what was legitimate' after the fact cost two weeks of archaeology through old reports.
Handled: Historic analytics and archived rank snapshots rebuilt the baseline while the disavow drafted in parallel, so classification and request preparation stopped blocking each other.
Obstacle
The first reconsideration request was rejected — the documentation of the process change was too thin to show the scheme could not simply resume.
Handled: The second request added the vendor correspondence, the full classification evidence, and a written link-acquisition policy; the manual action lifted in that second cycle.
Obstacle
The review cycle landed mid-way through the agency's biggest client launch, and the recovery kept competing for the same senior people.
Handled: Recovery decisions were batched into two fixed 45-minute weekly slots; client work kept its calendar and the request still went in on schedule.
The headline: manual action lifted in the second review cycle; organic entrances back to 92% of the pre-penalty baseline within 4 months — Removed, read from Search Console manual-actions report + analytics. A second check: referring domains earned from the research pieces at +11.
The agency now documents every link it earns and turns away vendors selling placements — the policy that lifted the penalty became the default. The team stopped flinching at Search Console notifications, and pitch prep meetings no longer include an apology for the agency's own rankings. The two research pieces keep feeding the new-business pipeline, which is the quiet irony everyone at the agency appreciates.
The result was read from Search Console manual-actions report + analytics against the pre-engagement baseline over the stated window, with a guardrail check on referring domains earned from the research pieces. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.
What we would do differently
We would have documented the earned-link baseline before touching the disavow file — reconstructing 'what was legitimate' after the fact cost two weeks.
[ Related service ]
[ Related builds ]
4 27Weekly prompt-set answers citing or summarizing the brand's content, weeks 1 to 10
14 31Qualified inquiries per month attributed to service pages, over the 8 weeks after rollout versus the 8 before
[ Next step ]
Next case study