How Google Detects Low-Quality Sites
Patent US9002832B1 explains how Google scores a site's incoming link profile to classify it as low quality — and why link diversity beats link volume.
What you'll learn
- Understand how Google's "link quality score" classifies sites as low quality
- Learn what diversity filtering and boilerplate filtering remove before scoring
- Know why many links from one domain barely move the needle
- Apply the patent's logic to avoid low-quality link patterns
The Problem Google Is Solving
Patent US9,002,832 B1 describes a method for classifying an entire site as "low quality" based on the distribution of quality among its incoming links. If a site's calculated link quality score falls below a threshold, its ranking is demoted for results pointing to that site.
Two Filters Applied Before Scoring
- Diversity filtering: multiple links from the same site, domain, or IP are collapsed into a single representative link (usually the highest-quality one). Links that share the same "content context" are also collapsed.
- Boilerplate filtering: links that only appear in navigation, sidebars, or footers (not in the main editorial content) are discarded entirely.
Only what survives both filters is used to calculate the site's link quality score — typically a ratio or weighted sum of links from "vital/good" quality groups versus "bad" ones.
100 links from a single donor site collapse into 1 representative link after diversity filtering. The total pool of "distinct" links barely grows, so the site's quality ratio doesn't meaningfully improve. Independent, varied donors matter far more than link volume from one source.
Practical Implications for Link Building
| Pattern | Effect under this patent |
|---|---|
| Many links from one PBN/network | Collapsed to ~1 representative link — minimal quality gain |
| Sitewide footer/nav links | Discarded entirely by boilerplate filtering |
| Independent, editorial, varied-context links | Each counted distinctly — directly raises the quality ratio |
| High share of low-quality ("bad") donors | Drags the site below the low-quality threshold |
The strategic conclusion: prioritize breadth of independent, editorial-context donors over raw link count, and avoid link patterns that only exist in boilerplate placements.
Key Takeaways
- Google can classify an entire site as "low quality" based on its incoming link profile, independent of on-page factors
- Diversity filtering collapses same-domain/IP/content-context links into one representative link before scoring
- Boilerplate-only links (nav, footer, sidebar) are excluded from the calculation entirely
- Mass links from one network or domain have a much smaller effect than the raw count suggests
- Prioritize independent, editorial, contextually varied donors over volume from a single source