Patents filed by search engine companies that describe how crawling, indexing, link analysis, and canonical selection actually work. Documented without interpretation layered on top.
A lot of technical SEO advice is based on observation, testing, and inference. That's valuable. But patents filed by search engine companies describe the actual systems — or at least the systems as they were designed at the time of filing. Reading the primary source is different from reading someone's interpretation of it.
Patents don't tell you everything. They describe systems that may have been implemented differently, or not at all. They may be outdated. But they give you a more grounded basis for understanding why certain technical issues matter and how crawlers are designed to handle them.
This library documents patents relevant to the topics covered in the explainers section. Each entry includes the patent number, filing date, and a plain-English summary of what the patent describes.
Google Inc.
The foundational patent describing the PageRank algorithm. Describes how the importance of a node (page) in a link graph is determined by the importance of the nodes that link to it, weighted by the number of outbound links from each linking node. Directly relevant to understanding why internal link structure affects how authority flows through a site.
Google Inc.
Describes a system for prioritizing which URLs a crawler visits based on signals including the importance of pages that link to the URL, the freshness of content, and the historical behavior of the server. Relevant to understanding why pages with no internal links receive less crawl attention than well-linked pages.
Google Inc.
Describes a method for determining how frequently a crawler should revisit a URL based on how often the content at that URL has changed historically. Pages that change frequently are scheduled for more frequent recrawling. Pages that rarely change are visited less often. Relevant to understanding how content freshness signals interact with crawl frequency.
Google Inc.
Describes methods for identifying URLs that return the same or substantially similar content. Relevant to understanding how search engines handle redirect chains that ultimately lead to duplicate content, and why having multiple URLs resolve to the same content without canonical signals creates indexing ambiguity.
Google Inc.
Describes a system for determining the canonical URL for a piece of content when multiple URLs return the same or equivalent content. Factors considered include the rel=canonical tag, the sitemap, the URL structure, and the internal link patterns pointing to each URL. Directly relevant to how redirect chains affect canonical selection.
Google Inc.
Describes how crawler systems use sitemap files to discover and prioritize URLs for crawling. Explains the relationship between sitemap signals and link-based discovery, and how conflicts between the two are handled. Relevant to understanding why sitemap errors — URLs in the sitemap that return non-200 status codes — create conflicting signals for the crawler.
Google Inc.
Describes a system that uses the lastmod date in sitemap files as a signal for scheduling recrawls. Also describes how the priority and changefreq attributes in sitemaps interact with other freshness signals. Relevant to understanding what information in your sitemap file actually influences crawl behavior and what is largely ignored.
Google Inc.
Describes a method for propagating signals — including quality scores and topic relevance — through the link graph from one page to related pages. Explains how a page's signals can influence the perceived quality of pages it links to. Relevant to understanding why the internal link structure of a site affects how individual pages are evaluated.
Google Inc.
Describes how the text of a hyperlink (anchor text) is used as a signal about the content of the linked page. Explains how anchor text from multiple sources is aggregated and weighted. Relevant to understanding why the text you use in internal links matters, not just the existence of the link itself.
Patents describe inventions as they were conceived at the time of filing. They do not necessarily describe systems as they are currently implemented. Search engine systems evolve continuously, and the gap between a patent's filing date and current implementation can be significant.
We document these patents as reference material — starting points for understanding the underlying concepts rather than definitive descriptions of current behavior. When a patent's described system aligns with observable behavior, that alignment is worth noting. When it doesn't, that's also worth noting.