The spokesman for Google to users, Danny Sullivan, said in a note blog how autocomplete predictions of the Google search engine were generated. Anyone Bahrain Email List who has ever searched the Google engine already has an idea of ​​what autocompletion is. Indeed, when entering the first letters of your request, predictions appear even before you have validated it. The predictions displayed as you type your query are meant to help save users time by giving them the information they need before they even complete their search. As Google explained in a section of the previous blog, those are not predictions and suggestions! Indeed, autocompletion is set up to help users complete a search they were about to do, and not to suggest new types of queries. How are these predictions generated?

The predictions that appear in the search bar are automatically generated based on real queries typed by Internet users into Google. 4 factors When generating these predictions, Google relies on 4 factors: Google therefore looks first at trending searches. This is the very first step in generating the predictions. ex autocompletion 2 Language and localization But of course, to offer relevant predictions adapted to each individual, Google will add two other factors, which are the user’s language and geographic location. In the example below, for the query “driving test”, which in French means “exam de license de rire”, you can see that the proposed predictions are different for a user located in California (on the left) and a user located in Ontario, Canada (right).

Don’t hesitate to use it for your SEO strategy!

ex autocompletion 3 As explained above, the freshness of the information is also taken into consideration to refine the predictions. If Google’s automated systems identify growing interest in a topic, they will display that prediction even if it is unrelated to user searches. For example, for basketball and NBA enthusiasts, it is more common to perform a search that includes the name of your favorite team, rather than looking for a specific game. However, if your team has just won a very important match, Google will base itself on the freshness of the information and show you adapted predictions. This is what the example below shows. By typing “Lakers” (for the NBA Los Angeles Lakers team),


Google offers us a prediction relating to a burning information, that is the victory in Game 6 of the 2020 NBA Finals against the Miami Heat, synonymous with the title. NBA for Lebron James and his cronies. Sometimes it happens that users make long search engine queries . In order to provide relevant predictions for these long queries, Google’s systems will analyze only part of the query, and not all of it. For example, to provide predictions for the long query below that is “name of the thing at the front of”, Google identified a lot of searches typed by Internet users who include “the front of” like “the front of a ship” or “the front of a boat”. It is therefore on this part of the request only that Google relied on to provide predictions.

And when there are no predictions?

Indeed, Google’s systems have more difficulty in identifying searches relating to “name of the thing” (in French “nom de la chose”). ex autocompletion 4 And when there are no predictions? Google has put measures in place to prevent unexpected predictions that may shock Internet users or lead to false information from appearing. For this, Google works in two ways: Automatic: the systems are made to prevent the appearance of these predictions which violate Google’s policy. Manual: If automatic systems fail to prevent these predictions, a dedicated team will delete them manually. Note also that the autocompletion will not display predictions considered by Google as violent, dangerous or sexual in nature.

Leave a Reply

Your email address will not be published.