When a journal vanishes from the index, the cause is always in the evidence.
Evidence in aggregate · figures from separate engagements
Metrics reflect real work across independent projects and do not represent a single engagement or identify any publisher.
The problem with ordinary SEO help
Most indexing services are built to launch a new journal. That is a different problem from a publisher who had coverage and lost it — after a migration, a redesign, or an attack.
Recovery is an investigation, not a checklist. It means reading crawl logs, backlink history, and metadata against each other until the cause is proven — then fixing it without breaking a single DOI. This works the same whether your articles live in WordPress, Joomla, OJS, hand-built HTML, or a custom stack.
Selected case files · findings, anonymized
Search engines were crawling ghost URLs left by the old platform instead of the articles. Mapped every crawlable path, redirected ~75,000 legacy URLs, cleaned the duplicate layers in staged cycles — every DOI target protected throughout.
A sudden flood of low-quality referring domains raised the question every publisher fears. Classified the profile against defined toxicity criteria, isolated the attack cohort, and filed a disavow with documented reasoning behind every entry.
301s resolved, yet Scholar coverage fell — because citation_* tags diverged between HTML and PDF during the move. Restored metadata parity, re-established full-text crawlability, tracked recovery against a dated baseline.
The method
Every case runs the same five phases — an evidence chain, not a guess.
Fix the exact date coverage dropped and everything that changed around it — migration, redesign, plugin, server move. Chain of custody starts here.
Crawl logs, Search Console, backlink history, a dated Google Scholar baseline, metadata across HTML and PDF. Nothing taken on faith.
Cross-reference the sources: attack vs update, shadow URLs vs metadata. The goal is causation, not correlation.
A written report with the diagnosis, the evidence behind it, and a stated confidence level. This is the deliverable.
DOI-safe fixes applied through staging, then monitoring so the same failure can't quietly return.
Types of investigation
Five kinds of case. Most publishers arrive with one — and discover they have another.
Why a collection disappeared from Google Scholar, Scopus, or Search — and how to bring it back.
Distinguish a real attack from an algorithm update with evidence, then contain it.
What a platform move actually broke — redirects, canonicals, metadata — and how to reverse the damage.
Detect AI-fabricated and hallucinated references before they reach publication.
Identify scrapers, prove how duplicate content cannibalizes the original, and remove it.
A law background alongside the technical work — rare in this field, and the reason the evidence holds up.
Standards & references · what the investigation is measured against
Every finding is checked against the sources that actually govern indexing — not opinion.
The technical rules a journal site must meet to be crawled and indexed. The baseline for every deindexation post-mortem.
Publisher-side policy: version grouping, primary-version selection, and how to request inclusion.
The Public Knowledge Project’s own troubleshooting reference for metadata and crawl failures.
The open-access indexing standard and Seal criteria that shape discoverability.
The editorial-integrity standards behind citation-fraud analysis.
DOI registration and the metadata a migration must never break.
The first 48 hours matter
Tell me the drop, the date, and the platform. You'll get a first read — attack, migration, or metadata — with evidence, not a guess.
Request an emergency diagnosis Paid diagnostic · written findings · confidence level stated