How Do Search Engines Decide Which Information is Relevant for The User?
Today it is about understanding how search engines decide which information is relevant for a user and how the search results are calculated after a search query. If possible, I don’t want to favor search engines in my series. But since Google is probably the most well-known in the world, I will explain many examples, again and again, using Google.
In the first part, we have already learned that search engines use crawlers to continuously collect information and websites from the entire public Internet, catalog it and store it in the search engine index. The data is already prepared so that the search engine can later work optimally when processing a search query (query processing). The Google search index contains billions of web pages and is over 100,000,000 gigabytes in size. It works like the index at the end of a book.
Based on What Information Does Google Generate Search Results?
Today, however, Google does not only collect and process words. Other data is also collected, indexed, and processed, such as the website’s server location, responsiveness, website structure, and loading speed. Prices, flight routes, job advertisements, and hotel rooms are also recorded and displayed in a structured manner in the search results.
However, Google does not only generate search results based on the data and relevance of a website. The location of the user performing the search, the device from which he does it (e.g. smartphone), and the way he does it (question? Purchase intention. Search for information) also play a significant role in the whole.
Which Ranking Factors Does a Search Engine Use When Answering a Search Query?
This question is not that easy to answer in general since nobody knows the weighting of the individual ranking factors and the Google algorithms exactly. But there are empirical observations, correlation studies, and SEO best practices that Google itself discloses. So today we know what is important for the optimization of a website and how Google reads data, interprets it, and presents it in the search results. The way Google processes and evaluates data is also constantly changing. Alone in the year 2018 there were over 3200 changes to the various Google algorithms. Anyone who has ever been to one of the 15 Google data centers worldwide will get a rough idea of the amount of data the search engine giant is dealing with.
“But Why is All This for Nothing?"
It’s simple: Google makes most of its revenue by showing ads. You can place advertisements for certain keywords in the Google search. If a user clicks on an ad, Google earns money and the advertiser may have a new customer. Since Google lives from a large number of users, the Google search must have added value, work free of manipulation, and be user oriented. No advertising revenue without users. Incidentally, the advertising system of Google “Google Ads” is strictly separated from the organic search. You can’t buy organic placement; you can only place an ad for keywords.
Which Method Is More Advantageous? Organic Growth or Pay for Ads?
Of course, getting a website to rank high organically (not paid) is much more cost-effective in the long run than paying for advertising. That’s why SEO is so important.
The Google search as it is today no longer a static keyword file, but a complex mechanism paired with artificial intelligence, machine learning, and quality assurance, which is intended to provide the user with the best possible search result. And that in over 170 languages and even artificial languages.
Important: I must again emphasize here that the informal scope in this series is kept to a minimum. If I were to explain all the features, functions, updates, and algorithms here, it would be like a novel.
How Does Google Generate The Rankings for The Search Results Pages?
- 1. Query Processing (The processing and answering of the search query)
- 2. Parsing (breaking down and converting the search query into usable components for the query processor)
How Search Terms are Converted by Google?
If you enter a search term or a question in the web interface of the Google search (e.g. Google.de), then these are first parsed after pressing the Enter key. This means, for example, that a question is only converted into meaningful and processable nouns (keywords), verbs (costs, means, weighs) and question words (what, who, how, where). Using these words, the query processor can then use the index and algorithms to determine the relevance of cataloged Calculate websites from the database and display them graphically in the Google web interface in less than a second. Basically, the parser is an interface to the database. Because Google maintains data centers worldwide, the proximity of a server with a parser and query processor is always very small, which enables high speed when answering a search query.
Problems Can be Addressed by Core Updates and BERT
The difficulty or rather the challenge for Google is to infer the intention behind a search phrase or a question as best as possible. That was also the main work of the developers in recent years. With Core Updates and updates like BERT, these challenges and problems have been addressed.
For example, the nouns provide information about the topic of the search and the verbs questions about the intention. Algorithms then decide with which weighting factor is applied to the database (index) and the query processor ultimately determines the order of the search results. The cataloging and sorting of the data are already done during the crawling.
So, we can see that the Google search is quite complex, and it takes several steps to get the search result. And at the same time, relevant advertisements from the Google Ads system are also imported. And all this in less than a second. Respect!
How Does Google Decide What Information is Relevant to the User?
- Does a website fit the search query thematically? (Relevance)
- Does a website match the search intent of the search query? (Intention)
- Does the search term appear on the website? (Relevance)
- Is the website accessible to the user? (Responsiveness on mobile devices)
- Is the quality of the website content high or even low? (e.g. copied or stolen)
- Is the author of a blog post known and recognized? (Content Authority)
- Is the website a well-known major newspaper or a niche blogger? (Authority)
- Is the website a well-known major newspaper or a niche blogger? (Authority)
- Is the information on a website linguistically and factually at the level of the search query? (academic / amateur)
Of course, these questions are only symbolic of the complexity behind them. I’m just trying to clarify how relevance can be checked. It’s important to understand that a website’s relevancy score for a keyword root or topic is already set in the index. A search query is almost always just “queried”. This also explains the secretly high speed with which Google can answer a search query.
If a user searches “transaction-oriented”, ie he wants to buy something, then online shops and shopping ads are more in the foreground. If a user searches in an “information-oriented” way, ie he is looking for a definition, for example, then Wikipedia entries, encyclopedias and wikis are more likely to appear. You know this yourself from your daily Google searches.
Search Engine and Ranking FAQ
This happens in under a second. Various factors and algorithms are used to generate the best possible search result or a search result page from the indexed data. Google always tries to answer the search query in the best possible way. This also applies to advertisements in Google search. The focus is always on the user and his search intention.
