Every day, millions of people conduct searches on Google, overwhelming the world’s most popular search engine. Somehow, Google’s search results return resources a user is likely looking for, regardless of the ambiguous nature of the keyword typed into the search box. Google uses a complex algorithm and many tools to return relevant searches that best fit the intent of the searcher. Those tools transform the user’s search terms into something the user may not recognize before returning search results that actually answer the query. SEO professionals can learn a lot about this process from Google’s own patent, and this knowledge may influence how content is created for the purposes of SEO. The following information will be presented in two parts.
Relevance and Search Volume in Google Queries
In the world of search engines, users often type in vague queries to the search box and expect miracles. In many cases, those miracles are happening, returning search results that somehow managed to answer the original query. Google is designed to handle queries that are too vague, too broad, too specific or even those that lack context. Their patent on providing query refinements can offer more details on the process, but basically, what is happening behind the scenes is a multi-step process
First, Google retrieves Web resources from its index based on associations it has with the query and keyword(s) the user typed in. Its algorithm next analyzes these resources for semantic clusters (topics, in a sense); in other words, it roughly groups and sorts those resources into broad conceptual categories. If those conceptual categories number too many, Google may provide some background refinement to better answer whatever is being asked of it by a user.
The next step Google takes is accessing something it calls an “association database”; this is the search engine’s repository for past search queries, Web resources, and the associations it has made between the two. For each association, the search engine will assign a degree of relevance, or “weight” to the query and multiply that weight figure to the search volume of the query. Search volume and frequency are interchangeable in this context. The scores it comes up with, coupled with past queries and the associations it has stored in its database, are analyzed again. The highest scoring semantic clusters and associations will become those Google presents to the user as a refined search result. As an example, think about what might happen if you typed the word “Trump” into Google’s search box. The word has several meanings, but the overwhelming majority of returned search results display information about Donald J. Trump, the 45th President of the United States. It is the search volume of the word “Trump” and the associations Google has made in the past that cause the search engine to refine what would otherwise be a very ambiguous query. The end result is that the user is presented with resources he or she could actually use to answer a question, rather than a mishmash of unrelated results.
In part two, we’ll talk about how context and location can also play a part in Google’s refinement of search queries. We will also talk about how SEO professionals can take advantage of Google’s patent information and details in developing content for clients. Continue reading on to part 2.