How Google Uses User Feedback to Rank Websites by Quality
A method for Google to judge a website's quality by looking at how users interact with its search results across different categories of search queries.
Patent Number
US 8615514
Status
Active
Filing Date
February 3, 2010
Grant Date
December 24, 2013
Expiration
February 3, 2030
Claims
36
Assignee
Google LLC
Inventors
Michael Flaster, Neil C. Fernandes, April R. Lehman
Citations
26 forward · 230 backward
What it covers
This patent describes a way to measure the quality of a website by grouping search data into specific buckets. It takes pairs of documents and the queries that led people to them, then sorts these pairs into partitions based on how well the document matched the query. By looking at how users click on results within these different buckets, the system calculates a 'skew' or statistical distribution of feedback. If a website performs unexpectedly across these partitions, it indicates a specific level of quality, which is then used as a signal to rank that website higher or lower in future search results.
What it doesn't cover
- —Does not cover ranking websites based on simple, non-partitioned metrics like total click count alone.
- —Does not cover manual human review of websites for quality assessment.
- —Does not cover ranking methods that ignore the relationship between the query and the specific document retrieved.
- —Does not cover systems that do not use statistical skew as a primary indicator of quality.
The clever bit
The innovation is using 'skew'—a measure of how lopsided a data distribution is—to detect quality. By comparing how a site performs on high-relevance vs. low-relevance queries, the system can identify if a site is genuinely useful or just gaming the system.
Why it matters
This patent is part of the technical foundation for modern search engine optimization (SEO) and ranking algorithms. It helps search engines distinguish between high-quality content and 'spammy' sites that might get clicks but don't actually satisfy user intent. It is a core component of how Google maintains the relevance of its search results at scale.
Real-world examples
- 1.Google Search ranking algorithms
- 2.Automated website quality scoring systems
- 3.Search engine result page (SERP) relevance tuning
Generated by PatentBrief · Not legal advice · patentbrief.org
US 8615514 · 2026