Google Update: Has Google switched to a new ranking algorithm?

Google Update: Has Google switched to a new ranking algorithm


Last week, Google started rolling out the “Core Updates March.” Already at the end of last week, there were some changes in the search results pages. It can be assumed that the update is still not completed.

Now there are suspicions that the core update could be due to the change to a new kind of ranking. This assumption is expressed at least in a recent contribution by the Search Engine Journal. The paper refers to a recent Google research paper entitled “Learning Groupwise Scoring Functions Using Deep Neural Networks” (PDF). It describes how the ranking of documents in groups can be done by comparing them with each other. Previous approaches evaluate documents only point by point concerning the individual document.

The summary of the document states:

However, they are restricted to pointwise scoring functions, i.e., the relevance score of a document is computed based on the document even, in this paper, we find this limitation by proposing generalized GroupWise scoring functions (GSFs), in which groups of documents in the list determine the relevance score of a document “

The age of the documents is also used for the evaluation, to close the information needs of the users. For example, if all of the documents in a search-relevant group refer to a topic in the past, then it can be concluded that in a specific case, users are interested in a past event rather than an actual event :

Calvin Harris is recent; the user may be interested in the latest news or tour information. On the other hand, it is more likely the user wants to learn about artist discography or biography. “

Google has repeatedly emphasized in the past that research paper approaches do not necessarily have to be used in the search. However, the illustrated methods for ranking web pages without specific use by Google for webmasters and SEOs can be helpful: they show that looking at similar web pages can be helpful to understand what is interesting for the users.

It is also clear that the ranking is less and less dependent on individual criteria such as the title or the meta-description. Rather, it is about viewing a website as a whole.