$ cat System Overview
Map of search and AI manipulation techniques, the signals they target, current policy context and the distinction between evidence and inference.
This overview maps policy-defined search manipulation, AI retrieval risks and malicious delivery. The March 2024 spam-policy changes and May 2024 documentation leak are historical context; current policy also includes August 2026 site-reputation changes. The guide distinguishes primary documentation, secondary analysis and proposed research rather than treating them as interchangeable evidence.
1 · What this archive covers
The archive contains 11 technique chapters, a lab library, case records, timeline, glossary, bibliography and changelog. The build indexes 48 tool catalogue records and counts 408 external-link occurrences; these are not counts of independently verified products or unique supporting sources. Chapter depth and evidence completeness vary. Proposed live studies are unexecuted; no original ranking experiments are reported.
2 · The shift that defines 2026
For two decades, manipulation targeted two things: links (votes) and content (relevance). The 2024 events pushed the frontier toward a third: engagement and provenance.
Engagement and AI retrieval add research questions; they do not prove that older tactics stopped working or that new manipulations succeed. Public 2023 court evidence concerning clicks also predates the documentation leak; see CTR/NavBoost.
3 · The technique map
The archive groups techniques by their target. Status labels are editorial categories, not measured prevalence or efficacy rates.
| Manipulates | Techniques | Status trend |
|---|---|---|
| Links / authority | PBNs, expired domains, link schemes (search manipulation) | Declining |
| Content at scale | Scaled content, affiliate exploitation | Active, heavily policed |
| Host trust | Parasite SEO / site reputation abuse | Declining (post-crackdown) |
| Engagement | CTR / NavBoost | Emerging interest, contested |
| Local signals | Local SEO exploitation | Specified civil review-rule prohibitions; platform policies also apply |
| AI answers | GEO manipulation, prompt injection | Emerging |
| Malicious delivery | SEO poisoning | Active (cybercrime) |
| Channel-pressured verticals | Restricted verticals (pharma, adult, alcohol) | Active; documented per vertical |
4 · Google's side of the arms race
The defensive systems are worth naming accurately, because a lot of black-hat writing describes them wrongly:
- SpamBrain — Google’s AI-based spam-prevention system; public descriptions do not reveal every production classifier or threshold. confirmed
- Helpful-content signals — originally a separate "Helpful Content System," but as of the March 2024 core update these signals were integrated into Google's core ranking systems, so they are no longer a distinct standalone system. confirmed
- Manual actions — human review can result in actions documented through Search Console. The archive supplies no comparison proving highest precision. Site-reputation effects now vary by searcher region. confirmed
5 · On numbers and honesty
Unsupported market, success, lifespan and detection figures have been withdrawn. A price or statistic requires a traceable source, scope, unit and date; an estimate badge cannot supply missing evidence. Fictional examples are labeled hypothetical and are not measurements. Build dates indicate generation time, not a fresh factual verification of every page. See Methodology.
6 · Where to go next
New here? Read Methodology for the evidence and safety framework, then PBNs for the full chapter format. Defending a site? Start with Detection & Defense. Curious about the AI frontier? See AI/GEO and prompt injection. Want the whole index? Open Find.
Primary context: Google ranking systems · August 2026 site-reputation update · FTC review-rule scope and effective date.