How the Google Search engine works

Google is a search engine created in 1996 by Stanford students Larry Page and Sergey Brin who originally published this project on a subdomain of their university's website — google.stanford.edu.

In the fall of 1997, the creators of the search engine officially registered a separate domain which later became the most visited website in the world — Google.com:

On September 4, 1998, Google Inc. was registered with the help of investors.

At first, the system was called "Backrub", then the name was changed to Google as this name sounds similar to the google number which is 10100. This name was chosen to express the desire to quickly streamline and systematize the huge amount of information published on the Internet.

The developers put the concept of PageRank at the heart of the service, according to which the importance of a web page from the point of view of the search engine was determined based on the number of links to it.

In late summer 1998, Larry Page and Sergey Brin received $100 thousand from the co-founder of Sun Microsystem for the development of the project. It was with these funds that their company was founded.

Since 2000, Google has launched paid advertisements that appear in search results for queries defined by the advertiser. The company expanded rapidly and gained popularity. In 2006, the verb "to google" became so common that it was even added to the Oxford Dictionary.

As of 2019, Google according to Forbes ranks second in the list of the most expensive brands in the world. This is the undisputed leader among search engines owning 92.42% of the global market share. According to Alexa statistics, the average user views more than 10 pages per day on Google, spending about 8 minutes on the website

In many cases, millions of pages match a user's search query. In order to streamline the display of information and show the most appropriate results, ranking principles are used.

Google's algorithms are based on the sequential execution of several processes related to each other which are crawling, page indexing, and then displaying them taking into account relevance and personalization.

Now, in addition to displaying relevant web pages, Google search allows retrieving information from books stored in the largest libraries, find out transport schedules, well-known facts, and much more. Such opportunities appeared thanks to the construction of the Knowledge Graph:

Webpage scanning is also called crawling — this is a process of identifying new pages on the Internet and updating information about those that are already indexed performed by Google search robots. Crawling is based on the Sitemap which is a file created for search engines containing all pages of a website.

During crawling, pages of the highest level are primarily processed since it is assumed that they are the most significant on a website. Then the Googlebot gradually moves to lower levels.

The Googlebot takes the website settings into account and processes those pages and links that are allowed for crawling. However, even if the ban on indexing a specific page is specified in the robots.txt file, it can still get into Google's search results. Therefore, if you want to reliably disallow pages from crawling, it is preferable to add the noindex attribute to the page HTML code or to write the noindex header in the HTTP request.

The frequency of crawling by a Google bot is determined by it independently; the process takes from several days to several weeks. You can request repeated crawling for individual pages or the entire website.

Google indexes pages after crawling. When crawled by Google robots, they systemize content by keywords and its newness creating a search index based on this data. It includes hundreds of billions of web pages spanning over one hundred million gigabytes.

Google can index content in almost any format:

Pages are added to the Google index if they follow webmaster guidelines.

You can speed up your website's indexing time on Google using the "URL inspection" tool in the Search Console. To do this, type the required URL and check whether it can be indexed:

After sending a request for indexing, in some cases, you can see the page in the Google search in 15 minutes. To do this, enter the request in this format:

For example:

You can also find out which pages of the website are indexed using the Search Console by going to the "Index" and then "Coverage":

Link ranking in Google depends on many factors. Experts believe that more than two hundred parameters affect the order in which Google displays search results. The PageRank algorithm patented by the creators of Google remains the ranking basis which determines the importance of pages depending on the number and quality of links to them.

The list of the most important ranking factors in 2019 includes:

• domain age and trust rate;
  
• quality of content;
  
• click-through rate (CTR) in organic results;
  
• adaptivity to mobile devices;
  
• page loading speed;
  
• search engine optimization of the page, that is, keyword entries, uniqueness, text volume, keywords in meta tags, etc.
  

Results for the same query may vary for different users. In order to take the individual needs of a particular user into account, Google uses search personalization based on their location, behavior, and social factors.

Geo-dependent queries take into account the geographic location of a user which is set by the browser, based on their IP address or using geolocation on a mobile device:

User behavior such as bookmarking pages and search history also affects the display of results. If a user has recently been looking for repair information, then if they type "black paint" as a query, they will see results related to construction:

If a user is not interested in repairs, but actively browses websites related to subjects of style and beauty, then in their personalized results there will be resources with hair dyes:

Personalization can also raise content in search results that social friends share with a user. You can manage this parameter using the search settings:

Here you can enable or disable private results: