An integral part of our lives, something we use every day; but how does Google search actually work?
According to Smart Insights, 81% of people search for a product or service online. With 70.38% of the search engine market share, Google is a huge household name. It enables us to research, educate and inform ourselves every day, but have you ever thought about how Google works and provides us with this information?
With thousands, sometimes millions of web pages to display, how does Google work in deciding which results to show and in which order?
Google themselves have kindly shared the ins and outs of their search engine, and we’re here to break it all down for you.
How Google Organises Web Content
Imagine all of the webpages online stored in a virtual library; this is essentially the Search Index. If this was a real library, it would no doubt be bigger than any library currently in existence; potentially even bigger than all libraries put together!
Before we even type out a new search, Google has organised information about the webpages within the Search Index.
Crawling to discover information
Google begins crawling web addresses from previous crawls and from sitemaps provided by website owners. Web crawlers go from page to page, collecting data about these sites and any changes; for example, new data, new page links or changes, and gives the information to Google to store on its servers.
Google Search Console allows website owners to monitor, maintain and make choices about how Google crawls their site and processes webpages. Through the console, owners can request a recrawl or opt out of crawling by using a robot.txt file.
Organising information by indexing
As crawlers find webpages, Google systems scan the content of the page in the same way a browser does. It records any key signals; including keywords and freshness and records it in the Search Index.
The Google Search Index contains an entry for every word on every webpage; each indexed webpage is added to the entries for all of the words it contains.
Google’s Knowledge Graph aims to better understand the people, places and things that searches care about, and goes beyond keyword matching. It does this by organising information not only from webpages, but also from other information sources too.
This includes searching text from libraries, finding travel times from public transport providers and navigating data from public sources like the World Bank.
How Google Matches Your Search Results
Almost instantly, using a highly guarded search algorithms, Google sifts through billions of web pages in our Search Index in a fraction of a second.
Its aim is to find the most relevant and useful results around the topic you have search for. Without some form of sorting, it would be nearly impossible to find what you need with all of the information available on the web.
A series of algorithms help Google to rank and sort results, to present the most suitable to the searcher.
These search algorithms look at lots of factors to provide suitable results. These include the words you have searched for, relevance and usability of pages, expertise of sources, your location and your settings.
The weight applied to these factors depends on the nature of the search; for example, freshness of content is more important for searches around news topics than it is for definition searches.
A rigorous process involving live tests and with Search Quality Raters from around the world are used. These ensure that the Search algorithms are a high standard and produce relevant results. Search Raters follow strict and lengthy guidelines that define the Search algorithm goals.
Key factors influencing search results:
Meaning of your query
Understanding intent of the searcher is the first thing Google needs to know. Language models understand what strings of words Google should look up in the Search Index.
Factors like spelling mistakes and type of query are deciphered at this point. Research on natural language and a synonym system match searches to pages describing the same thing. Developing this system took Google over 5 years, yet it improves results in over 30% of searches across different languages.
Categories of queries are also recognised. For example, words such as ‘review’, ‘pictures’, ‘opening hours’ or ‘how to’ are all recognised as a need for particular information. If a search is written in a certain language, the results displayed are likely to also be in that language.
Another important factor when determining the meaning of the query, is whether the user is seeking fresh content. Google’s freshness algorithms use trending keywords as signals for when more up-to-date content is more relevant than older pages.
Relevance of webpages
Analysing webpages discovers whether the content contains information relevant to a search or not.
On a basic level to see if a page is relevant, is if the keywords on a webpage match those used in a search query. If they appear on the webpage, particularly in the headings or body of text, then the information on the page is likely to be relevant to the user.
In addition to basic keyword matching, aggregated and anonymised interaction data is used to assess if results are relevant. This data is transformed into signals which help Google’s machine-learned systems to better estimate relevancy of results.
Ranking useful pages
Google algorithms determine how useful a webpage is. They analyse hundreds of factors from freshness of content, to the number of times search terms appear and if the page has a good user experience, to discover the best information to offer.
Trustworthiness and authority are also evaluated by how valuable other searchers found the pages. If high value sources link to pages, it indicates that those pages are likely to be relatable and valuable too.
Unfortunately, there can be a lot of spam sites out there. However, Google algorithms are exist to identify and remove sites that violate Google webmaster guidelines.
Usability of webpages
The algorithms recognise whether sites appear correctly in different browsers. Are they suitable for use on all device types and sizes? Do page loading times work well for slower internet connections? Google will work hard to present webpages in results that are easy to use.
Google provides guidance to website owners on how to improve their site usability with tools such as PageSpeed Insight and Webpagetest.org. These show owners what adjustments to tweak in order to improve the experience for website visitors.
Context and settings
Location, past search history and search settings all influence what Google thinks you will find useful and relevant.
For example, a search for ‘football’ would likely show results around English football and football from your area. If you search for ‘Barcelona’ and have previously searched for ‘Barcelona vs Arsenal’, then the search may be personalised as it could assume you would prefer results focused on football than the city itself.
You can control what Search activity and data influences your search experience by heading over to your Google Account.
How Google Works to Display Results In Helpful Ways
Google knows that users want answers as quickly and easily as possible. Therefore, it aims to present results in useful formats to better service the information you are looking for.
For example, in the SERPs (Search Engine Results Pages), Google shows directions with maps, images, videos, clickable telephone links or email addresses. These are always evolving. Google is always evaluating better ways to show these results to users for a better search experience.
For example, if you search for a weather forecast, you would probably prefer the weather to show on the results page, rather than as links to weather pages to click onto.
Every day, 15% of queries typed into Google Search are new queries that Google has never seen before. Therefore, it is constantly learning and also leaning on search algorithms to provide the best results to searches that it has never seen before.
How Does Google Work With Offering Ad Space?
Businesses or advertisers can pay for Ad Space at the top of the Google search results page. These are clear to see as they have an ‘Ad’ label, so that users can identify between paid for, or organic search results.
The organic results cannot be paid for to get a better or higher placement in the list of results. The list of organic results are in order of usefulness, relevancy and authority relating to your search query.
By offering paid-for advertising space, it enables Google to provide the search service for free. Ads are still relevant to the search terms; so if there are no useful ads to show, then the searcher will only see the normal organic results.
Providing the best user experience
Search is always evolving; with different trends, usage patterns, queries and so on.
Google engineers review how search can be improved by continuously testing and conducting thousands of experiments. These experiments turn into new changes and developments to the way we search.
All of the potential changes are rigorously tested before implementing. If tests show that the results are less useful, then the changes are not adopted. External Search Quality Raters continuously check that websites in search results give searchers the answers that they require.
Google will work to evaluate the quality of results based on the expertise, authoritativeness and trustworthiness of the content. We’ve gone into more detail on how to improve webpages for each of these areas by using on-page SEO in a recent blog post.
Summary: How does Google work?
So… how does Google work? With lots of web crawls, insight, search algorithms, research, testing and a lot of complex development work! Hopefully this gives you a (somewhat!) simplified overview of how Google works. So the next time you type in your search, you can appreciate just how much work has gone into those results!