Google explains how the site works to deliver exactly what you're looking for

Appearing in the first three results in Google’s organic search is the dream of every professional who works on the internet. This noble positioning guarantees the most clicks and helps significantly inflate site hits, which can mean increased sales for e-commerce, greater authority for companies or more credibility for content producers.

  • Google search will help you find home office jobs
  • Google will change algorithm to make it harder to use pages to spread slander
  • Deep analytics will have priority in Google product searches
  • Reaching the top is a very complex mission, so much so that it even has a dedicated area of ​​knowledge:

    Search Engine Optimization

    (IF THE). In practice, it means producing quality content based on Google’s best practices rules to please the public and, consequently, earn some extra points with the platform’s algorithm.

    Understand the operation of the search is fundamental for users and professionals (Image: Reproduction/WaybackMachine)

    Although it discloses general guidelines, the search giant does not reveal point by point what impacts the positioning of sites. It was precisely to better understand issues related to the search engine that the


    spoke exclusively with Danny Sullivan, Google’s executive in the search industry and one of the world’s leading experts in search tools. After working for years as a journalist, the professional started to dedicate himself to these mechanisms and became an expert to the point of being hired by the company.

    Want to catch up on the best tech news of the day?

    Access and subscribe to our new channel on youtube, Canaltech News.

    Every day a summary of the main news of the tech world for you!

    The criteria and the search

    The criteria used by the algorithm to position sites take into account the company’s mission to organize information and keep it accessible in a simple way. It may seem like an obvious statement, but it’s a lot more complicated than it appears. “This means going beyond the web, because there is a lot of information that is not on the internet”, explains Sullivan.

    The specialist separates the data organized by the search engine into five categories:

  • Information from the open web:

    websites, videos, images, news;

  • Common knowledge and public data:

    statistics, patents, academic studies, Wikipedia;

  • Partnerships:

    weather, scores sports, social media feeds, lyrics and finance;

  • Business :

    data submitted by entrepreneurs about businesses, locations and updates;

  • New information and insights:

    Street View, books, places of flooding, burning mapping;

    All these slopes are distinct from each other and need treatment, so the challenge, in the researcher’s view, is to understand the user’s research. “Imagine you’re walking down the street and someone says ‘pancakes.’ You will think: ‘ok, but what do you mean by that?’ And then it will ask some other questions to understand”, he contextualizes.

    Google, however, does not have the ability to return the user specific questions to understand what he wants to find, so he will do associations to deliver the results you consider most relevant.

    As it cannot ask the user questions, Google draws conclusions by itself ( Image: Danny Sullivan/Google)

    What can influence the results?

    To understand what you want to find, the Google algorithm considers these factors:

  • Subject current (the newest will be highlighted)
  • Your location
  • Previous Searches
  • Other words in the search
  • Language
  • The place where is f eita the search is an important factor for the result, because there are similar words with different meanings depending on the language. A bribe in Portuguese means bribery, but in Spanish it’s a tip, which will definitely result in different sites at the top of the search.

    There are also the famous synonyms, which drive any algorithm crazy if not considered the context of the sentence. Sullivan cites as an example the English word

    change, which is used as a substitute for terms such as adjust, convert, exchange, install, modify, replace and exchange money.

    “The challenge is to understand how all these words together create a meaning, a concept. It was based on this that BERT came about”, he emphasizes. BERT is a mechanism for understanding the so-called “natural language” of people, whose focus is not only to show results according to the terms, but rather to try to understand what the person meant.

    In the image below, you can see an example of how the system works. Before, the result pointed to a website that taught how to get a prescription. With BERT’s change, the answer became news that resolved the original doubt: whether you could pick up a drug at the pharmacy for a relative or friend.

    BERT takes literality results to a broader concept (Image: Danny Sullivan/Google)

    How does the search algorithm work?

    Sullivan explains that Google uses multiple signals in each search to associate your keyword with the results. This entire set of elements is analyzed in a fraction of a second to return sites that could possibly satisfy your wish.

    The first of these are the search terms. “If the page has the word, then it’s probably relevant, but you’ll have thousands of them in the same situation. If it appears in the title, it is even more likely”, he contextualizes. However, this alone would be too little to define the relevance, otherwise it would only be enough to repeat the word you want to rank several times on the page.

    The results are presented in different ways according to the term, according to what is most useful for the type of information you are looking for. That’s why searching for “Pizza” will show one result and “Pizza de pepperoni” will have a different one: in the first one, the searcher will understand that it is the concept of pizza, while in the second it could mean that you want to eat that flavor.

    Keywords are just one of the criteria used by Google (Image: Danny Sullivan/Google)

    Importance of links

    Following the words contextualized, other signals are also analyzed, but there is a very important factor: links. Most people only remember hyperlinks when they’re broken or don’t point to the expected result, but they’re fundamental on the web.

    “Think of links as recommendations: when you need a doctor , you ask trusted people who they refer to in the real world. On the internet, if there are many websites with links to a page, it may be more relevant than others”, explains the expert.

    Despite this, just having many links does not mean being a page trustworthy, in fact, since you need to know their origin to guarantee authority. Getting a referral from a highly reputable site like the New York Times, for example, is much more relevant to Google than your brother’s blog link created two hours ago.

    Does advertising influence?

    Sullivan explains that selling advertisements has no impact on the effective results of organic search. Google ADS markets keywords that put sites in the spotlight in the results, but doesn’t affect the search algorithm.

    “We sell advertising and display it when it can be relevant to people . We do not sell search history or information about what people are looking for. No one will do better in organic search results just because they paid for it,” he explains.

    Search ratings

    To ensure that the results are effective, the search engine has evaluators that allow you to say if it is adequate or if it needs improvement . “We have more than 10 thousand reviewers worldwide and it is important to have this diversity, because the search results are displayed for the entire planet. There are several people, from various countries, with different languages ​​so that we can be sure that we do the best job for everyone who uses the services”, he concludes.

    Qualified reviewers around the world analyze the search results (Image: Danny Sullivan/Google)

    These people work on the basis of a document of almost 70 pages with general guidelines on search quality. This is how everyone has the same parameters and uses similar structures when measuring results.

    According to Sullivan, evaluators do not monitor the sites themselves, individually, because this it would give rise to subjective criteria of favoring one over the other. “There are always new sites, so you’ll never see any raters looking at the results page by page, as it changes all the time. For this overall assessment you need automated systems like the ones used. What the evaluators do is analyze whether the result satisfies the search”, he adds.

    In practice, Google uses the information collected from these reliable sources, similar to the attitude of people when evaluating a restaurant or review the opinion of others before buying a product online. This data is gathered in the database and used to improve the system as a whole.

    More delicate topics require different treatment

    Within the Google rating system, there are some issues that are handled with more caution by the search engine and by the evaluators: a concept called Your Money or Your Life

    (YMYL), which means Your Money or Your Life. Everything that has an impact on lives or finances undergoes different review criteria.

    When evaluating the results for the research “How to get financing”, there is a greater concern not to show results with data theft attempts or wrong information that would lead to financial loss.

    The appraisers of the Google noticed the need to create a hub with information about Covid-19 (Image: Alveni Lisboa/Canaltech)

    An example given by Sullivan is the results about the Covid vaccine-10. In the beginning, the site was limited to showing technical information and news about immunizing agents, but later it began to show places where doses were applied, data on those vaccinated and information from local health departments.

    In this case, the algorithm has been improved to give special treatment to an unusual situation involving people’s health. This perception was obtained thanks to the work of the evaluators, who had the perception to notice the need to improve the result for these specific terms.

    And the future of searches?

    Danny Sullivan said he doesn’t have much to say about the future of search beyond what has already been announced for the market. In general, trends point to an increasing use of features such as Google Lens, which allows you to use the cell phone camera to perform searches, and to the integration of search with augmented reality features.

  • The Google search has evolved a lot from 446917 until today (Image: Danny Sullivan/Google)

    BERT is also constantly being improved to impact not only text searches, but also voice searches with Duplex technology . The latter was even one of the highlights of Google I/O 446917, with a very impressive interaction between human and machine thanks to BERT’s AI.

    Some curious data about the Google search engine:

  • % of daily searches are for new terms
  • The site index occupies more than 70 million GB
  • If all were printed as a book and placed side by side side could be done trips to the Moon
  • Per year, Google performs: 660.89 quality tests, 12.386 traffic experiments go, 19.557 comparative experiments and 4.660 algorithm updates.
  • Search filter detects and slashes 10 billions of spammy pages per day

    Did you like this article?

    Subscribe your email on Canaltech to receive daily updates with the latest news from the world of technology.


  • Related Articles

    Back to top button