Last updated on
In 2012, Google filed a patent titled “Ranking Search Results,” which outlines how branded search queries could be used as a ranking factor. This patent describes the use of branded and navigational queries, along with the count of independent links, as potential ranking factors. Despite being over a decade old, this patent may still influence Google’s ranking algorithm.
At the time, the search marketing community misunderstood the patent, leading to the loss of the valuable insights it contained.
The patent, titled “Ranking Search Results,” focuses on an invention designed to enhance how search results are ranked. It outlines an algorithm that re-ranks web pages based on two key factors:
Sorting Factor 1: Number of independent inbound links.
This factor counts the links to a site that are independent of the site being ranked.
Sorting Factor 2: Number of branded and navigational search queries.
These queries, referred to as “reference queries” or “implied links,” include branded and navigational searches.
The algorithm uses the counts from both factors to adjust the rankings of web pages accordingly.
First, I want to acknowledge that in 2012, I didn’t know how to properly interpret patents. At the time, I was more focused on research papers and left patent analysis to others. So, when I say that the search marketing community misunderstood the patent, I include myself in that group.
The “Ranking Search Results” patent was published in 2012, a year after the release of the Panda Update—a content quality update named after one of its engineers, Navneet Panda. Panda developed questions that third-party quality raters used to evaluate web pages, and these ratings helped determine if algorithm changes effectively filtered out “content farm” content.
Navneet Panda is also a co-author of the “Ranking Search Results” patent. Many SEOs saw his name on the patent and quickly assumed it was connected to the Panda update.
However, this assumption is incorrect. The Panda update is an algorithm that uses a “classifier” to assess web pages based on content quality. In contrast, the “Ranking Search Results” patent focuses solely on ranking search results and does not involve content quality or a content quality classifier.
There is no connection between the “Ranking Search Results” patent and the Panda update.
In 2009, Google introduced the Caffeine Update, which significantly improved its ability to quickly index fresh content. However, this update inadvertently created a loophole that allowed content farms to rank millions of web pages on topics that were rarely searched.
Former Google search engineer Matt Cutts explained the issue with content farms in an interview with Wired:
“It was like, ‘What’s the bare minimum that I can do that’s not spam?’ It sort of fell between our respective groups. And then we decided, okay, we’ve got to come together and figure out how to address this.”
Google’s response was the Panda Update, named after a search engineer who helped develop the algorithm specifically designed to filter out content farm material. To achieve this, Google used third-party site quality raters to evaluate websites, and the feedback was instrumental in defining content quality standards that targeted content farms.
Matt Cutts described the development process:
“There was an engineer who came up with a rigorous set of questions, everything from, ‘Do you consider this site to be authoritative? Would it be okay if this was in a magazine? Does this site have excessive ads?’ Questions along those lines. …we actually came up with a classifier to say, okay, IRS or Wikipedia or New York Times is over on this side, and the low-quality sites are over on this side. And you can really see mathematical reasons…”
In simple terms, a classifier is an algorithm within a system that categorizes data. For the Panda Update, the classifier categorized web pages based on their content quality.
When reading the “Ranking Search Results” patent, it’s evident that it does not address content quality but focuses solely on the ranking of search results.
The “Ranking Search Results” patent introduces two types of links that influence the modification of ranked search results:
Implied Links:
These are calculated from branded search queries and navigational queries, which the patent treats as links, referring to them as “implied links.” These implied links help generate a factor that adjusts the ranking of web pages relevant to search queries.
Express Links:
These are the traditional, independent inbound links to a web page. The patent uses these links in another calculation to create a factor that modifies the rankings of web pages responsive to a search query.
Both types of links—implied and express—are used as factors to adjust the rankings of a group of web pages.
The patent is relatively easy to understand, as it explains these concepts in straightforward language. It also uses specific terminology:
Here’s a key excerpt from the patent that outlines these concepts:
The second key part of the patent uses the same terminology to explain what implied links are:
Here’s how the patent describes implied links:
“A query can be classified as referring to a particular resource if the query includes a term that the system recognizes as referring to that resource.
For instance, a term may refer to a resource if it matches all or part of the resource’s identifier, such as its URL. For example, the term ‘example.com’ might be recognized as referring to the homepage of that domain, which has the URL ‘http://www.example.com.’
Therefore, search queries that include the term ‘example.com’ can be classified as referring to that homepage.
Similarly, if the system has data indicating that terms like ‘example sf’ and ‘esf’ are commonly used by users to refer to the resource with the URL ‘http://www.sf.example.com,’ then queries containing ‘example sf’ or ‘esf,’ such as ‘example sf news’ or ‘esf restaurant reviews,’ would be counted as reference queries for the group that includes the resource with that URL.”
In this explanation, “reference queries” are defined as terms that users commonly use to refer to a specific website. For example, if people search for “Walmart” alongside “Air Conditioner,” then the query “Walmart + Air Conditioner” is considered a “reference query” for Walmart.com, effectively counting as a citation and an implied link to that site.
Some SEOs mistakenly believe that a mention of a brand on a web page is treated by Google as if it were a link. They have misunderstood this patent, thinking it supports the idea that an “implied link” refers to a brand mention on a web page.
However, as the patent clearly outlines, “implied links” refer to references to brands within search queries, not on web pages. The concept is specific to how branded or navigational search queries are used, not about mentions of brands on sites.
The patent also addresses navigational queries:
“In addition or alternatively, a query can be categorized as referring to a particular resource when the query is determined to be a navigational query to that resource. From the user’s perspective, a navigational query is one submitted to reach a specific website or web page of a particular entity. The system can determine if a query is navigational to a resource by accessing data that identifies queries classified as navigational for various resources.”
The key takeaway is that the patent discusses using “reference queries” (branded or navigational search queries) as a factor similar to links, which is why they are termed “implied links.”
The algorithm described in the patent generates a “modification factor” that re-ranks a group of web pages relevant to a search query based on “reference queries” (which are branded search queries) and a count of independent inbound links.
Here’s how the modification or ranking process works:
To clarify, “resources” in this context refers to web pages and websites.
The patent explains the ranking process as follows:
“The system generates a modification factor for the group of resources from the count of independent links and the count of reference queries… For example, the modification factor can be a ratio of the number of independent links for the group to the number of reference queries for the group.”
Essentially, the patent filters out links associated with the website itself, focusing on independent links and branded search queries, which are then used as ranking factors (modification factors).
In hindsight, it was a mistake for some in the SEO industry to interpret this patent as “proof” that brand mentions on websites are a ranking factor.
It’s evident that “implied links” are not about brand mentions on web pages but rather about brand mentions (along with URLs and domains) in search queries, which can be used as ranking factors.
This patent outlines a method for using branded search queries as a signal of popularity and relevance when ranking web pages. It’s an effective signal because it reflects user behavior—indicating that users themselves find a specific website relevant for certain search queries. This type of signal is difficult to manipulate, making it a potentially reliable and non-spammy indicator.
While we don’t know for certain if Google employs the methods described in this patent, it’s easy to see why such a signal could still be valuable today.
Patents are written in specific, technical language, which can easily lead to misinterpretation if the context is ignored. A common mistake SEOs make is isolating one or two sentences from the broader context and using them to make assumptions about Google’s practices. This approach often leads to SEO misinformation.
To avoid such misunderstandings, it’s important to read patents carefully and within their full context. I recommend reading my article on How to Read Google Patents to learn how to interpret them correctly. Even if you don’t regularly read patents, understanding the basics can help you identify misinformation, which is unfortunately common.
In this article, I’ve focused on explaining what the “Ranking Search Results” patent is about and highlighting its key points. There are many detailed aspects and different implementations within the patent that I haven’t covered, as they aren’t essential for grasping the overall concept.
If you’re interested in those finer details, I strongly suggest reading my article on how to read patents before diving into the patent itself.
Original news from SearchEngineJournal
